Version bump

Copyright year update
Tests: Fix last commit on width (#4781 ).
2026-10-03 08:33:21 +02:00 · 2024-01-01 03:42:12 -05:00 · 2024-01-01 03:19:59 -05:00 · 2023-12-31 13:28:49 -05:00 · 2023-12-31 11:42:06 -05:00 · 2023-12-31 11:35:11 -05:00
2105 changed files with 177353 additions and 52539 deletions
@@ -63,6 +63,8 @@ ForEachMacros:
 SortIncludes:    true
 IncludeBlocks:   Regroup
 IncludeCategories:
+  - Regex:           '"V3Pch.*\.h"'
+    Priority:        -2 # Precompiled headers
  - Regex:           '"(config_build|verilated_config|verilatedos)\.h"'
    Priority:        -1 # Sepecials before main header
  - Regex:           '(<|")verilated.*'
@@ -112,6 +114,9 @@ SpacesBeforeTrailingComments: 2
 SpacesInAngles:  false
 SpacesInContainerLiterals: true
 SpacesInCStyleCastParentheses: false
+SpacesInLineCommentPrefix:
+  Minimum: 0
+  Maximum: -1
 SpacesInParentheses: false
 SpacesInSquareBrackets: false
 Standard:        Cpp11
@@ -0,0 +1,8 @@
+{
+	"name": "Verilator Build Environment",
+
+	"build": {
+        "dockerfile": "../ci/docker/buildenv/Dockerfile"
+    }
+
+}
@@ -11,6 +11,8 @@ Thanks for taking the time to report this.

 Can you attach an example that shows the issue?  (Must be openly licensed, ideally in test_regress format.)

+What 'verilator' command line do we use to run your example?
+
 What 'verilator --version' are you using?  Did you try it with the git master version?

 What OS and distribution are you using?
@@ -23,6 +23,10 @@ defaults:
    shell: bash
    working-directory: repo

+concurrency:
+  group: ${{ github.workflow }}-${{ github.event_name == 'pull_request' && github.ref || github.run_id }}
+  cancel-in-progress: true
+
 jobs:

  build:
@@ -33,15 +37,18 @@ jobs:
        compiler:
          - { cc: clang, cxx: clang++ }
          - { cc: gcc,   cxx: g++     }
-        m32: [0, 1]
+        # m32 1 is deprecated, not here to speed up regressions
+        m32: [0]
        exclude:
          # Build pull requests only with ubuntu-22.04 and without m32
-          - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-18.04' || 'do-not-exclude' }}
-          - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-20.04' || 'do-not-exclude' }}
-          - m32: ${{ github.event_name == 'pull_request' && 1              || 'do-not-exclude' }}
-          # Build -m32 only on ubuntu-22.04
+          # - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-18.04' || 'do-not-exclude' }}
+          # - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-20.04' || 'do-not-exclude' }}
+          # - m32: ${{ github.event_name == 'pull_request' && 1              || 'do-not-exclude' }}
+          # Build -m32 only on ubuntu-22.04 clang++
          - {os: ubuntu-18.04, m32: 1}
          - {os: ubuntu-20.04, m32: 1}
+          - compiler: { cc: gcc, cxx: g++ }
+            m32: 1
        include:
          # Build GCC 10 on ubuntu-20.04
          - os: ubuntu-20.04
@@ -56,7 +63,7 @@ jobs:
      CC: ${{ matrix.compiler.cc }}
      CXX: ${{ matrix.compiler.cxx }}
      CACHE_BASE_KEY: build-${{ matrix.os }}-${{ matrix.compiler.cc }}-m32=${{ matrix.m32 }}
-      CCACHE_MAXSIZE: 250M # Per build matrix entry (2000M in total)
+      CCACHE_MAXSIZE: 1000M # Per build matrix entry (* 5 = 5000M in total)
      VERILATOR_ARCHIVE: verilator-${{ github.sha }}-${{ matrix.os }}-${{ matrix.compiler.cc }}${{ matrix.m32 && '-m32' || '' }}.tar.gz
    steps:

@@ -101,16 +108,18 @@ jobs:
        compiler:
          - { cc: clang, cxx: clang++ }
          - { cc: gcc,   cxx: g++     }
-        m32: [0, 1]
+        m32: [0]
        suite: [dist-vlt-0, dist-vlt-1, dist-vlt-2, vltmt-0, vltmt-1]
        exclude:
          # Build pull requests only with ubuntu-22.04 and without m32
-          - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-18.04' || 'do-not-exclude' }}
-          - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-20.04' || 'do-not-exclude' }}
-          - m32: ${{ github.event_name == 'pull_request' && 1              || 'do-not-exclude' }}
+          # - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-18.04' || 'do-not-exclude' }}
+          # - os:  ${{ github.event_name == 'pull_request' && 'ubuntu-20.04' || 'do-not-exclude' }}
+          # - m32: ${{ github.event_name == 'pull_request' && 1              || 'do-not-exclude' }}
          # Build -m32 only on ubuntu-22.04
          - {os: ubuntu-18.04, m32: 1}
          - {os: ubuntu-20.04, m32: 1}
+          - compiler: { cc: gcc, cxx: g++ }
+            m32: 1
        include:
          # Test with GCC 10 on ubuntu-20.04 without m32
          - {os: ubuntu-20.04, compiler: { cc: gcc-10, cxx: g++-10 }, m32: 0, suite: dist-vlt-0}
@@ -127,7 +136,7 @@ jobs:
      CC: ${{ matrix.compiler.cc }}
      CXX: ${{ matrix.compiler.cxx }}
      CACHE_BASE_KEY: test-${{ matrix.os }}-${{ matrix.compiler.cc }}-m32=${{ matrix.m32 }}-${{ matrix.suite }}
-      CCACHE_MAXSIZE: 64M # Per build matrix entry (2160M in total)
+      CCACHE_MAXSIZE: 100M # Per build per suite (* 5 * 5 = 2500M in total)
      VERILATOR_ARCHIVE: verilator-${{ github.sha }}-${{ matrix.os }}-${{ matrix.compiler.cc }}${{ matrix.m32 && '-m32' || '' }}.tar.gz
    steps:

@@ -158,3 +167,30 @@ jobs:
      env:
        TESTS: ${{ matrix.suite }}
      run: ./ci/ci-script.bash
+
+  lint-py:
+    runs-on: ubuntu-22.04
+    name: Lint Python
+    env:
+      CI_BUILD_STAGE_NAME: build
+      CI_RUNS_ON: ubuntu-22.04
+      CI_M32: 0
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+      with:
+        path: repo
+
+    - name: Install packages for build
+      run: ./ci/ci-install.bash
+
+    # We use specific version numbers, otherwise a Python package
+    # update may add a warning and break our build
+    - name: Install packages for lint
+      run: sudo pip3 install pylint==3.0.2 ruff==0.1.3 clang sphinx sphinx_rtd_theme sphinxcontrib-spelling breathe ruff
+
+    - name: Configure
+      run: autoconf && ./configure --enable-longtests --enable-ccwarn
+
+    - name: Lint
+      run: make -k lint-py
@@ -0,0 +1,85 @@
+# Build and push verilator docker image when tags are pushed to the repository.
+# The following variable(s) must be configured in the github repository:
+# DOCKER_HUB_NAMESPACE: docker hub namespace.
+# The following secrets must be configured in the github repository:
+# DOCKER_HUB_USER: user name for logging into docker hub
+# DOCKER_HUB_ACCESS_TOKEN: docker hub access token.
+name: Build Verilator Container
+
+on:
+  push:
+    tags: [ 'v*' ]
+  workflow_dispatch:
+    inputs:
+      manual_tag:
+        description: 'Git tag to use for image build'
+        required: true
+        type: string
+      add_latest_tag:
+        description: 'Tag workflow_dispatch docker image as "latest"'
+        required: true
+        type: boolean
+        default: false
+
+jobs:
+
+  build:
+
+    runs-on: ubuntu-22.04
+
+    strategy:
+      matrix:
+        contexts:
+          - "ci/docker/run:verilator"
+          # - "ci/docker/buildenv:verilator-buildenv"
+
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+
+    - name: Extract context variables
+      run: |
+        echo "${{ matrix.contexts }}" | sed -r 's/(.*):.*/build_context=\1/' >> "$GITHUB_ENV"
+        echo "${{ matrix.contexts }}" | sed -r 's/.*:(.*)/image_name=\1/' >> "$GITHUB_ENV"
+        echo "git_tag=${GITHUB_REF#refs/*/}" >> "$GITHUB_ENV"
+
+    - name: Use manual tag
+      if: ${{ inputs.manual_tag }}
+      run: |
+        echo "git_tag=${{ inputs.manual_tag }}" >> "$GITHUB_ENV"
+
+    - name: Docker meta
+      id: docker_meta
+      uses: docker/metadata-action@v4
+      with:
+        images: |
+          ${{ vars.DOCKER_HUB_NAMESPACE }}/${{ env.image_name }}
+        tags: |
+          type=match,pattern=(v.*),group=1,enable=${{ startsWith(github.ref, 'refs/tags/v') }}
+          type=raw,value=${{ inputs.manual_tag }},enable=${{ inputs.manual_tag != '' }}
+          type=raw,value=latest,enable=${{ inputs.add_latest_tag == true }}
+
+    - name: Set up QEMU
+      uses: docker/setup-qemu-action@v2
+
+    - name: Set up Docker Buildx
+      uses: docker/setup-buildx-action@v2
+      with:
+          buildkitd-flags: --debug
+
+    - name: Login to Docker Hub
+      uses: docker/login-action@v2
+      with:
+        username: ${{ secrets.DOCKER_HUB_USER }}
+        password: ${{ secrets.DOCKER_HUB_ACCESS_TOKEN }}
+
+    - name: Build and Push to Docker
+      uses: docker/build-push-action@v4
+      if: startsWith(github.ref, 'refs/tags/v')
+      with:
+        context: ${{ env.build_context }}
+        build-args: SOURCE_COMMIT=${{ env.git_tag }}
+        platforms: linux/arm64,linux/amd64
+        push: ${{ !env.ACT && startsWith(github.ref, 'refs/tags/v') }}
+        tags: ${{ steps.docker_meta.outputs.tags }}
+        labels: ${{ steps.docker_meta.outputs.labels }}
@@ -22,6 +22,10 @@ defaults:
  run:
    working-directory: repo

+concurrency:
+  group: ${{ github.workflow }}-${{ github.event_name == 'pull_request' && github.ref || github.run_id }}
+  cancel-in-progress: true
+
 jobs:

   windows:
@@ -41,3 +41,6 @@ verilator-config-version.cmake
 **/_build/*
 **/obj_dir/*
 /.vscode/
+/.idea/
+/cmake-build-*/
+/test_regress/snapshot/
@@ -4,7 +4,7 @@
 #
 #*****************************************************************************
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
@@ -15,7 +15,7 @@
 cmake_minimum_required(VERSION 3.15)
 cmake_policy(SET CMP0091 NEW) # Use MSVC_RUNTIME_LIBRARY to select the runtime
 project(Verilator
-    VERSION 5.003
+    VERSION 5.020
    HOMEPAGE_URL https://verilator.org
    LANGUAGES CXX
 )
@@ -23,7 +23,8 @@ project(Verilator
 option(DEBUG_AND_RELEASE_AND_COVERAGE
    "Builds both the debug and release binaries, overriding CMAKE_BUILD_TYPE. Not supported under MSBuild.")

-set(PYTHON3 python)
+find_package(Python3 COMPONENTS Interpreter)
+set(PYTHON3 Python3::Interpreter)
 set(CMAKE_INSTALL_DATADIR ${CMAKE_INSTALL_PREFIX})
 include(GNUInstallDirs)
 include(CMakePackageConfigHelpers)
@@ -34,7 +35,7 @@ if (NOT WIN32)
    message(WARNING "CMake support on Linux/OSX is experimental.")
 endif()

-if (MSVC)
+if (WIN32)
    if (DEFINED ENV{WIN_FLEX_BISON})
        set(WIN_FLEX_BISON "$ENV{WIN_FLEX_BISON}")
    endif()
@@ -44,14 +45,27 @@ if (MSVC)
    if (NOT WIN_FLEX_BISON)
        message(FATAL_ERROR "Please install https://github.com/lexxmark/winflexbison and set WIN_FLEX_BISON environment variable. Please use install cmake target after a successful build.")
    endif()
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /std:c++20")
-else()
-    if (NOT DEFINED ENV{FLEX_INCLUDE})
-        message(FATAL_ERROR "Please set environment variable FLEX_INCLUDE to the directory containing FlexLexer.h")
+    set(CMAKE_CXX_STANDARD 20)
+endif()
+
+set(OBJCACHE "" CACHE STRING "Path for ccache, auto-detected if empty")
+option(OBJCACHE_ENABLED "Compile Verilator with ccache" ON)
+
+if (OBJCACHE_ENABLED)
+    if (OBJCACHE STREQUAL "")
+        find_program(OBJCACHE_PATH ccache)
+        if (OBJCACHE_PATH STREQUAL "OBJCACHE_PATH-NOTFOUND")
+            set(OBJCACHE_PATH "")
+        endif()
+    else()
+        set(OBJCACHE_PATH "${OBJCACHE}")
    endif()
-    set(WIN_FLEX_BISON "$ENV{FLEX_INCLUDE}")
-    if (WIN32 AND NOT MINGW)
-        set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++20")
+    if (NOT OBJCACHE_PATH STREQUAL "")
+        execute_process(COMMAND "${OBJCACHE_PATH}" --version
+                OUTPUT_VARIABLE objcache_version)
+        string(REGEX MATCH "[^\n\r]+" objcache_version "${objcache_version}")
+        message(STATUS "Found ccache: ${OBJCACHE_PATH} (\"${objcache_version}\")")
+        set(CMAKE_CXX_COMPILER_LAUNCHER "${OBJCACHE_PATH}")
    endif()
 endif()

@@ -99,6 +113,7 @@ configure_package_config_file(verilator-config-version.cmake.in verilator-config
 install(FILES ${CMAKE_CURRENT_BINARY_DIR}/verilator-config-version.cmake DESTINATION ${CMAKE_INSTALL_PREFIX})

 foreach (program
+    verilator
    verilator_gantt
    verilator_ccache_report
    verilator_difftree
@@ -118,6 +133,7 @@ install(DIRECTORY include TYPE DATA FILES_MATCHING
    PATTERN "include/verilated.mk"
    PATTERN "include/*.[chv]"
    PATTERN "include/*.cpp"
+    PATTERN "include/*.sv"
    PATTERN "include/gtkwave/*.[chv]*"
    PATTERN "include/vltstd/*.[chv]*"
 )
@@ -8,7 +8,436 @@ The changes in each Verilator version are described below.  The
 contributors that suggested a given feature are shown in []. Thanks!


-Verilator 5.006 2022-01-22
+Verilator 5.020 2024-01-01
+==========================
+
+**Major:**
+
+* Support compilation with precompiled headers with Make and GCC or CLang.
+* Change include to systemc instead of systemc.h (#4622) (#4623). [Chih-Mao Chen]
+  This may require that SystemC programs add 'using namespace sc_core', 'using namespace sc_dt'.
+
+**Minor:**
+
+* Add devcontainer support (#4748). [Stefan Wallentowitz]
+* Support `iff` in sensitivity list (#1482) (#4626). [Krzysztof Bieganski, Antmicro Ltd.]
+* Support parameterized virtual interfaces (#4047) (#4743). [Ryszard Rozak, Antmicro Ltd.]
+* Support --timing triggers for virtual interfaces (#4673). [Krzysztof Bieganski, Antmicro Ltd.]
+* Support ccache when compiling Verilator with CMake (#4678). [Anthony Donlon]
+* Support passing constraints to --xml-only output (still otherwise unsupported) (#4683). [Shahid Ikram]
+* Support node memory usage information in --stats (#4684). [Geza Lore]
+* Support vpiConstType in vpi_get() (#4761). [Todd Strader]
+* Support vpi_iterate on packages with vpiInstance (#4726). [Todd Strader]
+* Support multiple parameters in virtual interfaces (#4745). [Ryszard Rozak, Antmicro Ltd.]
+* Support user C/C++ code in final archive, and make a lib{model}.a (#4749) (#4754). [Fan Shupei]
+* Support inside operator on unpacked arrays and queues (#4751). [Ryszard Rozak, Antmicro Ltd.]
+* Support VPI parameter iteration (#4765). [Todd Strader]
+* Support packages in vpi_handle_by_name() (#4768). [Todd Strader]
+* Support invoking interface methods on virtual interface variables (#4774) (#4775). [Jordan McConnon]
+* Remove deprecated options (#4663). [Geza Lore]
+* Remove older compiler support; require C++14 or newer (#4784) (#4786).
+* Optimize timing-delayed queue (#4584). [qrqiuren]
+* Optimize substitute optimization memory usage (#4687). [Geza Lore]
+* Optimize wide primitive operations with -Oz (#4733). [Geza Lore]
+* Optimize V3Premit performance etc. (#4736). [Geza Lore]
+* Fix VPI TOP level variable iteration (#3919) (#4618). [Marlon James]
+* Fix display with no % printing assoc array (#4376). [Alex Solomatnikov]
+* Fix scheduling of external force signals (#4577) (#4668). [Geza Lore]
+* Fix a memory leak in V3Fork (#4628). [Krzysztof Boroński]
+* Fix linking parameterized hierarchical blocks and recursive hierarchical blocks (#4654). [Anthony Donlon]
+* Fix identifiers that end with '_' on Windows (#4655). [Anthony Donlon]
+* Fix 'for' loop with outside variable reference (#4660). [David Harris]
+* Fix tracing FST enums (#4661) (#4756). [Todd Strader]
+* Fix interface parameters used in loop generate constructs (#4664) (#4665). [Anthony Donlon]
+* Fix C++20 compilation errors (#4670).
+* Fix deadlocks in error handler (#4672). [Mariusz Glebocki, Antmicro Ltd.]
+* Fix MingW compilation (#4675). [David Ledger]
+* Fix trace when using SystemC with certain configurations (#4676). [Anthony Donlon]
+* Fix range access to classes depending on parameter resolution (#4681). [Krzysztof Boroński]
+* Fix select into constant And/Or/Xor pattern (#4689). [Geza Lore]
+* Fix access type of function arguments (#4692) (#4694). [Ryszard Rozak, Antmicro Ltd.]
+* Fix dynamic NBAs with automatic vars (#4696). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix # 0 delays for process resumption, etc. (#4697). [Krzysztof Boroński]
+* Fix conflicted namespace for coroutines (#4701) (#4707). [Jinyan Xu]
+* Fix compilers seeing empty input due to file system races (#4708). [Flavien Solt]
+* Fix shift of > 32-bit number (#4719). [Flavien Solt]
+* Fix Windows include gates in filesystem Flush implementation. (#4720). [William D. Jones]
+* Fix power operator with wide numbers and constants (#4721) (#4763). [Flavien Solt]
+* Fix parameter passing to ports (#4723). [Ryszard Rozak, Antmicro Ltd.]
+* Fix block names of nested do..while loops (#4728). [Ryszard Rozak, Antmicro Ltd.]
+* Fix class name in error on 'new' on virtual class (#4739). [Ryszard Rozak, Antmicro Ltd.]
+* Fix typedefs pointing to parameterized classes (#4747). [Ryszard Rozak, Antmicro Ltd.]
+* Fix $finish twice to no longer exit (#4757). [Tim Hutt]
+* Fix dynamic NBA conditions (#4773). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix `V3Fork` stage to run only if `--timing` is set (#4778). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix max multiply width and add runtime assertions if too small. (#4781)
+
+
+Verilator 5.018 2023-10-30
+==========================
+
+**Major:**
+
+* Support compilation with precompiled headers with Make and GCC or CLang.
+* Change include of systemc instead of systemc.h (#4622) (#4623). [Chih-Mao Chen]
+  This may require that SystemC programs add 'using namespace sc_core', 'using namespace sc_dt'.
+
+**Minor:**
+
+* Add SIDEEFFECT warning on mishandled side effect cases.
+* Add trace() API even when Verilated without --trace (#4462). [phelter]
+* Add warning on interface instantiation without parens (#4094). [Gökçe Aydos]
+* Add sv_vpi_user.h from IEEE 1800-2017 Annex M (#4606). [Marlon James]
+* Support 'disable fork' (#4125) (#4569). [Aleksander Kiryk, Antmicro Ltd.]
+* Support 'wait fork' (#4586). [Aleksander Kiryk, Antmicro Ltd.]
+* Support 'randc' (#4349).
+* Support assigning events (#4403). [Krzysztof Boroński]
+* Support resizing function call inout arguments (#4467).
+* Support NBAs in non-inlined functions/tasks (#4496) (#4572). [Krzysztof Bieganski, Antmicro Ltd.]
+* Support converting parameters inside modules to localparams (#4511). [Anthony Donlon]
+* Support concatenation of unpacked arrays (#4558). [Yutetsu TAKATSUKASA]
+* Support Clang 16 (#4592). [Mariusz Glebocki]
+* Support VPI variables of real and string data types (#4594). [Marlon James]
+* Support making VL_LOCK_SPINS configurable (#4599). [Geza Lore]
+* Change code --stats output (#4597). [Geza Lore]
+* Change --prof-exec infrastructure and report (#4602). [Geza Lore]
+* Change lint_off to not propagate upwards to files including where the lint_off is.
+* Optimize empty expression statements (#4544).
+* Optimize trace internals (#4610) (#4612). [Geza Lore]
+* Optimize internal performance issues (#4638). [Geza Lore]
+* Fix conversion of impure logical expressions to bit expressions (#487 partial) (#4437). [Ryszard Rozak, Antmicro Ltd.]
+* Fix enum functions in localparams (#3999). [Andrew Nolte]
+* Fix passing arguments by reference (#3385 partial) (#4489). [Ryszard Rozak, Antmicro Ltd.]
+* Fix multithreading handling to separate by code units that use/never use it (#4228). [Mariusz Glebocki, Antmicro Ltd.]
+* Fix usage of annotation options (#4486) (#4504). [Michal Czyz]
+* Fix detecting local vars in nested forks (#4493) (#4506). [Kamil Rakoczy]
+* Fix handling input file path separator (#4515) (#4516). [Anthony Donlon]
+* Fix mis-support for parameterized UDPs (#4518). [Anthony Donlon]
+* Fix constant conversion of $realtobits, $bitstoreal (#4522).  [Andrew Nolte]
+* Fix conversion of integers in $display '%e' (#4528). [muzafferkal]
+* Fix non-inlined interface tracing (#3984) (#4530). [Todd Strader]
+* Fix stream operations with operands of struct type (#4531) (#4532). [Ryszard Rozak, Antmicro Ltd.]
+* Fix 'this' in a constructor (#4533). [Ryszard Rozak, Antmicro Ltd.]
+* Fix stream shift operator of 32 bits (#4536). [Julien Faucher]
+* Fix object destruction after a copy constructor (#4540) (#4541). [Ryszard Rozak, Antmicro Ltd.]
+* Fix inlining of real functions miscasting (#4543). [Andrew Nolte]
+* Fix broken link error for enum references (#4551). [Anthony Donlon]
+* Fix logical expressions with class objects - caching in v3Const (#4552). [Ryszard Rozak, Antmicro Ltd.]
+* Fix using functions/tasks following class definition inside module (#4553). [Anthony Donlon]
+* Fix large constant buffer overflow (#4556). [Varun Koyyalagunta]
+* Fix instance arrays connecting to array of structs (#4557). [raphmaster]
+* Fix error message for invalid parameter overrides (#4559). [Anthony Donlon]
+* Fix shift to remove operation side effects (#4563).
+* Fix compile warning on unused member function variable (#4567).
+* Fix method narrowing conversion compiler error (#4568).
+* Fix interface comparison (#4570). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix dynamic triggers for named events (#4571). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix dictionaries with keys of class types (#4576). [Ryszard Rozak, Antmicro Ltd.]
+* Fix to not remap local assign intervals in forks (#4583). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix display optimization ignoring side effects (#4585).
+* Fix PLI/DPI user defined system task/function grammar (#4587) (#4588). [Quentin Corradi]
+* Fix fault on empty clocking block (#4593). [Alex Mykyta]
+* Fix creating implicit nets for inputs of gate primitives (#4603). [Geza Lore]
+* Fix try_put method of unbounded mailbox (#4608). [Ryszard Rozak, Antmicro Ltd.]
+* Fix stable name generation in V3Fork (#4615) (#4624). [Krzysztof Boroński]
+* Fix virtual methods (#4616). [Ryszard Rozak, Antmicro Ltd.]
+* Fix insertion at queue end (#4619). [Krzysztof Boroński]
+* Fix rand fields of reference types (#4627). [Ryszard Rozak, Antmicro Ltd.]
+* Fix dynamic casts of null values (#4631). [Ryszard Rozak, Antmicro Ltd.]
+* Fix signals read via virtual interfaces being misoptimized (#4645). [Krzysztof Bieganski, Antmicro Ltd.]
+* Fix handling of static keyword in methods (#4649). [Ryszard Rozak, Antmicro Ltd.]
+* Fix preprocessor to show `line 2 on resumed file.
+
+
+Verilator 5.016 2023-09-16
+==========================
+
+**Minor:**
+
+* Add prepareClone and atClone APIs for Verilated models (#3503) (#4444). [Yinan Xu]
+* Add check for conflicting options e.g. binary and lint-only (#4409). [Ethan Sifferman]
+* Add --no-trace-top to not trace top signals (#4412) (#4422). [Frans Skarman]
+* Support recursive function calls (#3267).
+* Support assignments of packed values to stream expressions on queues (#4401). [Ryszard Rozak, Antmicro Ltd]
+* Support no-parentheses calls to static methods (#4432). [Krzysztof Boroński]
+* Support block_item_declaration in forks (#4455). [Krzysztof Boroński]
+* Support assignments of stream expressions on queues to packed values (#4458). [Ryszard Rozak, Antmicro Ltd]
+* Support function non-constant default arguments (#4470).
+* Support 'let'.
+* Optimize Verilator executable size by refactoring error reporting routines (#4446). [Anthony Donlon]
+* Optimize Verilation runtime pointers and graphs (#4396) (#4397) (#4398). [Krzysztof Bieganski, Antmicro Ltd]
+* Optimize preparations towards multithreaded Verilation (#4291) (#4463) (#4476) (#4477) (#4479).  [Kamil Rakoczy, Antmicro Ltd]
+* Fix Windows filename format, etc (#3873) (#4421). [Anthony Donlon].
+* Fix t_dist_cppstyle Perl performance issue (#4085). [Srinivasan Venkataramanan]
+* Fix using type in parameterized classes without #() (#4281) (#4440). [Anthony Donlon]
+* Fix false INFINITELOOP on forever..mailbox.get() (#4323). [Srinivasan Venkataramanan]
+* Fix data type of condition operation on class objects (#4345) (#4352). [Ryszard Rozak, Antmicro Ltd]
+* Fix variables mutated under fork..join_none/join_any blocks into anonymous objects (#4356). [Krzysztof Boroński]
+* Fix V3CUse, do not consider implementations (.cpp) at all (#4386). [Krzysztof Boroński]
+* Fix ++/-- under statements (#4399). [Aleksander Kiryk, Antmicro Ltd]
+* Fix detection of mixed blocking and nonblocking assignment in nested assignments (#4404). [Ryszard Rozak, Antmicro Ltd]
+* Fix jumping over object initialization (#4411). [Krzysztof Boroński]
+* Fix multiple issues towards short circuit support (#4413) (#4460). [Ryszard Rozak, Antmicro Ltd]
+* Fix variable lifetimes in extern methods (#4414). [Krzysztof Boroński]
+* Fix multiple function definitions in V3Sched (#4416). [Hennadii Chernyshchyk]
+* Fix false UNUSEDPARAM on generate localparam (#4427). [Bill Pringlemeir]
+* Fix checking for parameter and port connections in the wrong place (#4428). [Anthony Donlon]
+* Fix coroutine handle movement during queue manipulation (#4431). [Aleksander Kiryk, Antmicro Ltd]
+* Fix nested assignments on the LHS (#4435). [Ryszard Rozak, Antmicro Ltd]
+* Fix false MULTITOP on bound interfaces (#4438). [Alex Solomatnikov]
+* Fix internal error on real conversion (#4447). [vdhotre-ventana]
+* Fix lifetime unknown error on enum.name (#4448). [jwoutersymatra]
+* Fix unstable output of VHashSha256 (#4453). [Anthony Donlon]
+* Fix static cast from a stream type (#4469) (#4485). [Ryszard Rozak, Antmicro Ltd]
+* Fix error on enum with VARHIDDEN of cell (#4482). [Michail Rontionov]
+* Fix lint of case statements with enum and wildcard bits (#4464) (#4487). [Anthony Donlon]
+* Fix reference to extended class in parameterized class (#4466).
+* Fix recursive display causing segfault (#4480). [Kuoping Hsu]
+* Fix the error message when the type of ref argument is wrong (#4490). [Ryszard Rozak, Antmicro Ltd]
+* Fix display %x formatting of real.
+* Fix mis-warning on #() in classes' own functions.
+* Fix IGNOREDRETURN to not warn on void-cast static function calls.
+* Fix ZERODLY to not warn on 'wait(0)'.
+
+
+Verilator 5.014 2023-08-06
+==========================
+
+**Minor:**
+
+* Deprecation planned for 32-bit pointer -m32 mode (#4268).
+* Deprecate CMake config below version 3.13 (#4389) (#4390). [Vito Gamberini]
+* Support some stream operations on queues (#4292). [Ryszard Rozak, Antmicro Ltd]
+* Support property declaration with empty parentheses (#4313) (#4317). [Anthony Donlon]
+* Support locator methods with "with" on assoc arrays (#4335). [Ryszard Rozak, Antmicro Ltd]
+* Support string replication with variable (#4341). [Aleksander Kiryk, Antmicro Ltd]
+* Support more types in wait (#4374). [Aleksander Kiryk, Antmicro Ltd]
+* Support static method calls as default values of function arguments (#4378). [Ryszard Rozak, Antmicro Ltd]
+* Add GENUNNAMED lint warning. [Srinivasan Venkataramanan, Deepa Palaniappan]
+* Add MISINDENT lint warning for misleading indentation.
+* Fix 'VlForkSync' redeclaration (#4277). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix processes that can outlive their parents (#4253). [Krzysztof Boronski, Antmicro Ltd]
+* Fix duplicate fork names (#4295). [Ryszard Rozak, Antmicro Ltd]
+* Fix splitting coroutines (#4297) (#4307). [Jiamin Zhu]
+* Fix error when multiple duplicate DPI exports (#4301).
+* Fix class reference assignment checking (#4296). [Ryszard Rozak, Antmicro Ltd]
+* Fix handling of ref types in initial values of type parameters (#4304). [Ryszard Rozak, Antmicro Ltd]
+* Fix comparison of string parameters (#4308). [Ryszard Rozak, Antmicro Ltd]
+* Fix state update for always processes (#4311). [Aleksander Kiryk, Antmicro Ltd]
+* Fix multiple edge timing controls in class methods (#4318) (#4320) (#4344). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix implicit calls of base class constructors with optional arguments (#4319). [Ryszard Rozak, Antmicro Ltd]
+* Fix propagation of process requirement (#4321). [Krzysztof Boroński]
+* Fix unhandled overloads in V3InstrCount (#4324). [Krzysztof Boroński]
+* Fix selects of static members (#4326). [Ryszard Rozak, Antmicro Ltd]
+* Fix references to members of results of static methods (#4327). [Ryszard Rozak, Antmicro Ltd]
+* Fix unique..with method on queues of class objects (#4328). [Ryszard Rozak, Antmicro Ltd]
+* Fix queue slicing (#4329). [Aleksander Kiryk, Antmicro Ltd]
+* Fix wildcard referring types (#4336) (#4342). [Aleksander Kiryk, Antmicro Ltd]
+* Fix comparison of class objects (#4346). [Ryszard Rozak, Antmicro Ltd]
+* Fix unexpected RefDType on assoc arrays (#4337). [Aleksander Kiryk, Antmicro Ltd]
+* Fix cmake astgen for Rocky Linux 8.7 (#4343). [Julian Daube]
+* Fix class timescale in class packages (#4348). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix string concatenations (#4354). [Ryszard Rozak, Antmicro Ltd]
+* Fix unlinked task error from broken context (#4355) (#4402). [Aleksander Kiryk, Antmicro Ltd]
+* Fix selects on unpacked structs (#4359). [Ryszard Rozak, Antmicro Ltd]
+* Fix select operation on assoc array with wide keys (#4360). [Ryszard Rozak, Antmicro Ltd]
+* Fix non-public methods with wide output (#4364). [Ryszard Rozak, Antmicro Ltd]
+* Fix handling of super.new calls (#4366). [Ryszard Rozak, Antmicro Ltd]
+* Fix assign to input var in methods (#4367). [Aleksander Kiryk, Antmicro Ltd]
+* Fix VlProcess not found (#4368). [Aleksander Kiryk, Antmicro Ltd]
+* Fix order of evaluation of function calls in statements (#4375). [Ryszard Rozak, Antmicro Ltd]
+* Fix config_build.h issues (#4380) (#4381). [Andrew Miloradovsky]
+
+
+Verilator 5.012 2023-06-13
+==========================
+
+**Major:**
+
+* With -j or --build-jobs, multithread Verilator's emit phase of Verilation. [Kamil Rakoczy, Antmicro Ltd]
+  Additional Verilator-internal stages will become multithreaded over time.
+
+**Minor:**
+
+* Add --main-top-name option for C main TOP name (#4235) (#4249). [Don Williamson]
+* Add creating __inputs.vpp file with --debug (#4177). [Tudor Timi]
+* Add NEWERSTD warning when using feature in newer language standard (#4168) (#4172). [Ethan Sifferman]
+* Add warning that timing controls in DPI exports are unsupported (#4238). [Krzysztof Bieganski, Antmicro Ltd]
+* Support std::process class (#4212). [Aleksander Kiryk, Antmicro Ltd]
+* Support inside expressions with strings and doubles (#4138) (#4139). [Krzysztof Boroński]
+* Support get_randstate/set_randstate class method functions.
+* Support for condition operator on class objects (#4214). [Ryszard Rozak, Antmicro Ltd]
+* Support array max (#4275). [Aleksander Kiryk, Antmicro Ltd]
+* Optimize VPI callValueCbs (#4155). [Hennadii Chernyshchyk]
+* Configure for faster C++ linking using 'mold', if it is installed.
+* Fix crash on duplicate imported modules (#3231). [Robert Balas]
+* Fix false WIDTHEXPAND on array declarations (#3959). [JOTEGO]
+* Fix marking overridden methods as coroutines (#4120) (#4169). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix SystemC signal copy macro use (#4135). [Josep Sans]
+* Fix duplicate static names in blocks in functions (#4144) (#4160). [Stefan Wallentowitz]
+* Fix initialization order of initial static after function/task (#4159). [Kamil Rakoczy, Antmicro Ltd]
+* Fix linking AstRefDType if it has parameterized class ref (#4164) (#4170). [Ryszard Rozak, Antmicro Ltd]
+* Fix crash caused by $display() optimization (#4165) (#4166). [Tudor Timi]
+* Fix arrays of unpacked structs (#4173). [Risto Pejašinović]
+* Fix $fscanf of decimals overflowing variables (#4174). [Ahmed El-Mahmoudy]
+* Fix super.new missing data type (#4147). [Tudor Timi]
+* Fix missing class forward declarations (#4151). [Krzysztof Boroński]
+* Fix hashes of instances of parameterized classes (#4182). [Ryszard Rozak, Antmicro Ltd]
+* Fix forced assignments that override non-continuous assignments (#4183) (#4192). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix wide structure VL_TOSTRING_W generation (#4188) (#4189). [Aylon Chaim Porat]
+* Fix references to members of parameterized base classes (#4196). [Ryszard Rozak, Antmicro Ltd]
+* Fix tracing undefined alignment (#4201) (#4288) [John Wehle]
+* Fix class specific same methods for AstVarScope, AstVar, and AstScope (#4203) (#4250). [John Wehle]
+* Fix dotted references in parameterized classes (#4206). [Ryszard Rozak, Antmicro Ltd]
+* Fix bit selections under parameterized classes (#4210). [Ryszard Rozak, Antmicro Ltd]
+* Fix duplicate std:: declaration with -I (#4215). [Harald Pretl]
+* Fix deep traversal of class inheritance timing (#4216). [Krzysztof Boroński]
+* Fix class parameters of enum types (#4219). [Ryszard Rozak, Antmicro Ltd]
+* Fix static methods with prototypes (#4220). [Ryszard Rozak, Antmicro Ltd]
+* Fix LATCH warning on function local variables (#4221) (#4284) [Julien Margetts]
+* Fix VCD scope types (#4227) (#4282). [Àlex Torregrosa]
+* Fix incorrect multi-driven lint warning (#4231) (#4248). [Adrien Le Masle]
+* Fix missing assignment for wide unpacked structs (#4233). [Jiamin Zhu]
+* Fix unpacked struct == and != operators (#4234) (#4240). [Risto Pejašinović]
+* Fix AstStructSel clean when data type is structure (#4241) (#4244). [Risto Pejašinović]
+* Fix function calls in with statements (#4245). [Ryszard Rozak, Antmicro Ltd]
+* Fix operator == for unpacked struct, if elements are VlUnpacked arrays (#4247). [Risto Pejašinović]
+* Fix STATIC lifetime for variables created from clocking items (#4262). [Krzysztof Boroński]
+* Fix names of foreach blocks (#4264). [Ryszard Rozak, Antmicro Ltd]
+* Fix iterated variables in foreach loops to have VAUTOM lifetimes (#4265). [Krzysztof Boroński]
+* Fix missing assignment for wide class members (#4267). [Jiamin Zhu]
+* Fix the global uses timing flag when forks exist (#4274). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix struct redefinition (#4276). [Aleksander Kiryk, Antmicro Ltd]
+* Fix detection of wire/reg duplicates.
+* Fix false IMPLICITSTATIC on package functions.
+* Fix method calls on function return values.
+
+
+Verilator 5.010 2023-04-30
+==========================
+
+**Minor:**
+
+* Add --public-depth to force public to a certain instance depth (#3952). [Andrew Nolte]
+* Add --public-params flag (#3990). [Andrew Nolte]
+* Add CONSTRAINTIGN warning when constraint ignored.
+* Add STATICVAR warning and convert to automatic (#4018) (#4027) (#4030). [Ryszard Rozak, Antmicro Ltd]
+* Add error if class types don't match (#4064). [Ryszard Rozak, Antmicro Ltd]
+* Support class extends of package::class.
+* Support class srandom and class random stability.
+* Support class method calls without parenthesis (#3902) (#4082). [Srinivasan Venkataramanan]
+* Support method calls without parenthesis (#4034). [Ryszard Rozak, Antmicro Ltd]
+* Support parameterized return types of methods (#4122). [Ryszard Rozak, Antmicro Ltd]
+* Support parameterized class references in extends statement (#4146). [Ryszard Rozak, Antmicro Ltd]
+* Support complicated IEEE 'for' assignments.
+* Support $fopen as an expression.
+* Support ++/-- on dotted member variables.
+* Optimize static trigger evaluation (#4142). [Geza Lore, X-EPIC]
+* Optimize more xor trees (#4071). [Yutetsu TAKATSUKASA]
+* Change range order warning from LITENDIAN to ASCRANGE (#4010). [Iztok Jeras]
+* Change ZERODLY to a warning.
+* Fix random internal crashes (#666). [Dag Lem]
+* Fix install, standardization in cmake CMakeLists.txt (#3974). [Yu-Sheng Lin]
+* Fix UNDRIVEN warning seg fault (#3989). [Felix Neumärker]
+* Fix symbol entries when inheriting classes (#3995) (#3996). [Krzysztof Boroński]
+* Fix event controls reusing same variable (#4014). Kamil Rakoczy <[email protected]>
+* Fix push to dynamic queue in struct (#4015). [ezchi]
+* Fix names for blocks in do..while loop (#4019). [Ryszard Rozak, Antmicro Ltd]
+* Fix randomize on null field (#4023). [Ryszard Rozak, Antmicro Ltd]
+* Fix rand fields in base classes (#4025). [Ryszard Rozak, Antmicro Ltd]
+* Fix large return blocks with --comp-limit-blocks (#4028). [tenghtt]
+* Fix clocking block scope internal error (#4032). [Srinivasan Venkataramanan]
+* Fix false LATCH warning on --assert 'unique else if' (#4033) ($4054). [Jesse Taube]
+* Fix characters from DEFENV literals for Conda (#4035) (#4044). [Tim Snyder]
+* Fix info message prints under --assert (#4036) (#4053). [Srinivasan Venkataramanan]
+* Fix C++ compile errors when passing class refs as task argument (#4063). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix NBAs inside fork-joins (#4050). [Aleksander Kiryk, Antmicro Ltd]
+* Fix task calls as fork statements (#4055). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix _Vilp used before declaration (#4057) (#4062). [Josep Sans]
+* Fix incorrect optimization of bit op tree (#4059) (#4070). [Yutetsu TAKATSUKASA]
+* Fix parameters in a class body to be localparam (#4061). [Ryszard Rozak, Antmicro Ltd]
+* Fix interface generate begin (#4065). [Srinivasan Venkataramanan]
+* Fix tracing with awaits at end of block (#4075) (#4076). [Krzysztof Bieganski, Antmicro Ltd]
+* Fix sense expression variable naming (#4081). [Kamil Rakoczy]
+* Fix importing symbols from base class (#4084). [Ryszard Rozak, Antmicro Ltd]
+* Fix false error on new const assignment (#4098). [Tudor Timi]
+* Fix unpacked structs under classes (#4102). [Tudor Timi]
+* Fix variables in class methods to be automatic (#4111) (#4137). [Peter Monsson]
+* Fix to use parallel build for projects with a lot of files (#4116). [Krzysztof Boroński]
+* Fix including __Syms header in generated C++ files (#4123). [Krzysztof Boroński]
+* Fix systemc namespace issues (#4126) (#4127). [Eyck Jentzsch]
+* Fix class param extends A=B (#4128). [Ryszard Rozak, Antmicro Ltd]
+* Fix missing begin block hierarchy in --xml-only cells section (#4129) (#4133). [Risto Pejašinović]
+* Fix resolution of class lvalues after parameterization (#4131). [Krzysztof Boroński]
+* Fix DFG error on $countbits (#4101) (#4143). [Paul Donahue]
+* Fix duplicating parameter class types (#4115). [Ryszard Rozak, Antmicro Ltd]
+* Fix class extend param references (#4136).  [Ryszard Rozak, Antmicro Ltd]
+* Fix -CFLAGS to allow overriding optimization levels (#4140). [Peter Monsson]
+* Fix DPI function type alias (#4148) (#4149). [Toru Niina]
+* Fix deleting unused parameterized classes (#4150). [Ryszard Rozak, Antmicro Ltd]
+* Fix false ENUMVALUE on expressions and arrays.
+* Fix unnecessary verilated_std.sv waivers in --waiver-output.
+
+
+Verilator 5.008 2023-03-04
+==========================
+
+**Minor:**
+
+* Add --annotate-points option, change multipoint on line reporting (#3876). [Nassim Corteggiani]
+* Add --verilate-jobs option (#3889). [Kamil Rakoczy, Antmicro Ltd]
+* Add WIDTHEXPAND and WIDTHTRUNC warnings to replace WIDTH (#3900). [Andrew Nolte]
+* Add SOURCE_DATE_EPOCH for docs/guide/conf.py (#3918). [Larry Doolittle]
+* Add /*verilator public[flat|flat_rd|flat_rw| ]*/ metacomments (#3894). [Joseph Nwabueze]
+* Add lint warning on always_comb multidriven (#3888) (#3939). [Adam Bagley]
+* Add warning on ++/-- over expressions with potential side effects (#3976). [Krzysztof Boroński]
+* Add error on mixing .name and by-port instantiations.
+* Removed deprecated --cdc option.
+* Support unpacked unions.
+* Support interface classes and class implements.
+* Support global clocking and $global_clock.
+* Support class parameters without initial values.
+* Support cast to numbers from strings.
+* Support struct I/O in --lib-create (#3378) (#3892). [Varun Koyyalagunta]
+* Support function calls without parenthesis (#3903) (#3902). [Ryszard Rozak, Antmicro Ltd]
+* Support class extending its parameter (#3904). [Ryszard Rozak, Antmicro Ltd]
+* Support static function variables (#3830). [Ryszard Rozak, Antmicro Ltd]
+* Support recursive methods (#3987). [Ryszard Rozak, Antmicro Ltd]
+* Fix real parameters of infinity and NaN.
+* Fix pattern assignment to unpacked structs (#3510). [Mostafa Garnal]
+* Fix single-element replication to dynarray/unpacked/queue (#3548). [Gustav Svensk]
+* Fix constant enum methods (#3621). [Todd Strader]
+* Fix inconsistent naming of generate scope arrays (#3840). [Andrew Nolte]
+* Fix namespace fallback resolution (#3863) (#3942). [Aleksander Kiryk, Antmicro Ltd]
+* Fix std:: to be parsed first (#3864) (#3928). [Aleksander Kiryk, Antmicro Ltd]
+* Fix cmake warning if multiple SOURCES w/o PREFIX (#3916) (#3927). [Yoda Lee]
+* Fix parameterized class function linkage (#3917). [Ryszard Rozak]
+* Fix static members of type aliases of a parameterized class (#3922). [Ryszard Rozak, Antmicro Ltd]
+* Fix class extend parameter dot case (#3926). [Ryszard Rozak, Antmicro Ltd]
+* Fix MsWin missing directory exception, and ::std (#3928) (#3933) (#3935). [Kritik Bhimani]
+* Fix very long VPI signal names (#3929). [Marlon James]
+* Fix VPI upper interface scopes not found (#3937). [David Stanford]
+* Fix virus detection false positive (#3944). [Stuart Morris]
+* Fix constant string function assignment (#3945). [Todd Strader]
+* Fix constant format field widths (#3946). [Todd Strader]
+* Fix class field linking when a super classes is a param (#3949). [Ryszard Rozak, Antmicro Ltd]
+* Fix CMake bad C identifiers (#3948) (#3951). [Zixi Li]
+* Fix build on HP PA architecture (#3954). [John David Anglin]
+* Fix date on the front page of verilator.pdf (#3956) (#3957). [Larry Doolittle]
+* Fix associative arrays declared with ref type (#3960). [Ryszard Rozak, Antmicro Ltd]
+* Fix missing error on negative replicate (#3963). [Benjamin Menküc]
+* Fix self references to parameterized classes (#3962). [Ryszard Rozak, Antmicro Ltd]
+* Fix LITENDIAN warning is backwards (#3966) (#3967). [Cameron Kirk]
+* Fix subsequent parameter declarations (#3969). [Ryszard Rozak, Antmicro Ltd]
+* Fix timing delays to not truncate below 64 bits (#3973) (#3982). [Felix Neumärker]
+* Fix cmake on MacOS to mark weak symbols with -U linker flag (#3978) (#3979). [Peter Debacker]
+* Fix UNDRIVEN warning seg fault (#3989). [Felix Neumärker]
+* Fix coverage of class methods (#3998). [Tim Paine]
+* Fix packed array structure replication.
+* Fix enum.next(0) and enum.prev(0).
+
+
+Verilator 5.006 2023-01-22
 ==========================

 **Minor:**
@@ -20,7 +449,8 @@ Verilator 5.006 2022-01-22
 * Support property calls without parenthesis (#3879) (#3893). [Ryszard Rozak, Antmicro Ltd]
 * Support import/export lists in modport (#3886). [Gökçe Aydos]
 * Support class queue equality (#3895). [Ilya Barkov]
-* Add IMPLICITSTATIC warning when a ftask/function is implicitly static (#3839). [Ryszard Rozak, Antmicro Ltd]
+* Support type case and type equality comparisons.
+* Add IMPLICITSTATIC warning when a task/function is implicitly static (#3839). [Ryszard Rozak, Antmicro Ltd]
 * Add VL_VALUE_STRING_MAX_WORDS override (#3869). [Andrew Nolte]
 * Optimize expansion of extend operators.
 * Internal multithreading tests. [Mariusz Glebocki, et al, Antmicro Ltd]
@@ -34,7 +464,7 @@ Verilator 5.006 2022-01-22
 * Fix self references when param class instantiated (#3833). [Ryszard Rozak, Antmicro Ltd]
 * Fix memory leak in V3Sched, etc. (#3834). [Geza Lore]
 * Fix compatibility with musl libc / Alpine Linux (#3845). [Sören Tempel]
-* Fix empty case items crash (#3851). [rporter]
+* Fix empty case items crash (#3851). [Rich Porter]
 * Fix VL_CPU_RELAX on MIPS/Armel/s390/sparc (#3843) (#3891). [Kamil Rakoczy]
 * Fix module parameter name collision (#3854) (#3855). [James Shi]
 * Fix unpacked array expansion (#3861). [Joey Liu]
@@ -44,6 +474,8 @@ Verilator 5.006 2022-01-22
 * Fix foreach unnamedblk duplicate error (#3885). [Ilya Barkov]
 * Fix elaboration of member selected classes (#3890). [Ilya Barkov]
 * Fix mismatched widths in DFG (#3872). [Geza Lore, Yike Zhou]
+* Fix lint for non-integral types in packed structs.
+* Fix generate case with empty body statements.


 Verilator 5.004 2022-12-14
@@ -143,7 +575,7 @@ Verilator 5.002 2022-10-29
 * Fix null access on optimized-out fork statements (#3658). [Krzysztof Bieganski, Antmicro Ltd]
 * Fix VPI inline module naming mismatch (#3690) (#3694). [Jiuyang Liu]
 * Fix deadlock in timeprecision when using SystemC (#3707). [Kamil Rakoczy, Antmicro Ltd]
-* Fix width mismatch on inside operator (#3714). [Alex Torregrosa]
+* Fix width mismatch on inside operator (#3714). [Àlex Torregrosa]


 Verilator 4.228 2022-10-01
@@ -373,7 +805,7 @@ Verilator 4.214 2021-10-17
 * Fix removing if statement with side effect in condition (#3131). [Alexander Grobman]
 * Fix --waiver-output for multiline warnings (#2429) (#3141). [Keith Colbert]
 * Fix internal error on bad widths (#3140) (#3145). [Zhanglei Wang]
-* Fix crash on clang 12/13 (#3148). [Kouping Hsu]
+* Fix crash on clang 12/13 (#3148). [Kuoping Hsu]
 * Fix cygwin compile error due to missing -std=gnu++14 (#3149). [Sun Kim]
 * Fix $urandom_range when the range is 0 ... UINT_MAX (#3161). [Iru Cai]
 * Fix constructor-parameter argument comma-separation in C++ (#3162). [Matthew Ballance]
@@ -488,12 +920,12 @@ Verilator 4.202 2021-04-24
 * Mark --no-relative-cfuncs as scheduled for deprecation.
 * Add --coverage-max-width (#2853). [xuejiazidi]
 * Add VerilatedCovContext::forcePerInstance (#2793). [Kevin Laeufer]
-* Add FST SystemC tracing (#2806). [Alex Torregrosa]
+* Add FST SystemC tracing (#2806). [Àlex Torregrosa]
 * Add PINNOTFOUND warning in place of error (#2868). [Udi Finkelstein]
 * Support overlaps in priority case statements (#2864). [Rupert Swarbrick]
 * Support for null ports (#2875). [Udi Finkelstein]
 * Fix class unpacked-array compile error (#2774). [Iru Cai]
-* Fix scope types in FST and VCD traces (#2805). [Alex Torregrosa]
+* Fix scope types in FST and VCD traces (#2805). [Àlex Torregrosa]
 * Fix exceeding command-line ar limit (#2834). [Yinan Xu]
 * Fix false $dumpfile warning on model save (#2834). [Yinan Xu]
 * Fix --timescale-override not suppressing TIMESCALEMOD (#2838). [Kaleb Barrett]
@@ -550,7 +982,7 @@ Verilator 4.110 2021-02-25
 **Minor:**

 * Support concat selection (#2721).
-* Support struct scopes when dumping structs to VCD (#2776) [Alex Torregrosa]
+* Support struct scopes when dumping structs to VCD (#2776) [Àlex Torregrosa]
 * Generate SELRANGE for potentially unreachable code (#2625) (#2754) [Pierre-Henri Horrein]
 * For --flatten, override inlining of public and no_inline modules (#2761) [James Hanlon]
 * Fix little endian interface pin swizzling (#2475). [Don Owen]
@@ -561,7 +993,7 @@ Verilator 4.110 2021-02-25
 * Fix class extends with VM_PARALLEL_BUILDS (#2775). [Iru Cai]
 * Fix shifts by > 32 bit values (#2785). [qrq992]
 * Fix examples not flushing vcd (#2787). [Richard E George]
-* Fix little endian packed array pattern assignment (#2795). [Alex Torregrosa]
+* Fix little endian packed array pattern assignment (#2795). [Àlex Torregrosa]


 Verilator 4.108 2021-01-10
@@ -1037,7 +1469,7 @@ Verilator 4.014 2019-05-08
 **Minor:**

 * Add --trace-fst-thread.
-* Support '#' comments in $readmem. (#1411) [Frederick Requin]
+* Support '#' comments in $readmem. (#1411) [Frédéric Requin]
 * Support "'dx" constants. (#1423) [Udi Finkelstein]
 * For FST tracing use LZ4 compression. [Tony Bybell]
 * Add error when use parameters without value. (#1424) [Peter Gerst]
@@ -1307,19 +1739,19 @@ Verilator 3.912 2017-09-23

 **Minor:**

-* Support or/and/xor array intrinsic methods. (#1210) [Mike Popoloski]
+* Support or/and/xor array intrinsic methods. (#1210) [Michael Popoloski]
 * Support package export. (#1217) [Usuario Eda]
-* Support module port parameters without defaults. (#1213) [Mike Popoloski]
+* Support module port parameters without defaults. (#1213) [Michael Popoloski]
 * Add performance information to --stats file.
 * Simplify VL_CONST_W macro generation for faster compiles.
 * Optimize improvements for Shift-And, and replication constructs.
-* Fix ordering of arrayed cell wide connections. (#1202 partial) [Mike Popoloski]
-* Fix LITENDIAN warning on arrayed cells. (#1202) [Mike Popoloski]
-* Fix enum ranges without colons. (#1204) [Mike Popoloski]
-* Fix GCC noreturn compile error. (#1209) [Mike Popoloski]
-* Fix constant function default parameters. (#1211) [Mike Popoloski]
-* Fix non-colon array of interface modports. (#1212) [Mike Popoloski]
-* Fix .name connections on interfaces. (#1214) [Mike Popoloski]
+* Fix ordering of arrayed cell wide connections. (#1202 partial) [Michael Popoloski]
+* Fix LITENDIAN warning on arrayed cells. (#1202) [Michael Popoloski]
+* Fix enum ranges without colons. (#1204) [Michael Popoloski]
+* Fix GCC noreturn compile error. (#1209) [Michael Popoloski]
+* Fix constant function default parameters. (#1211) [Michael Popoloski]
+* Fix non-colon array of interface modports. (#1212) [Michael Popoloski]
+* Fix .name connections on interfaces. (#1214) [Michael Popoloski]
 * Fix wide array indices causing compile error.


@@ -1441,8 +1873,8 @@ Verilator 3.888 2016-10-14

 **Minor:**

-* Add --no-decoration to remove output comments, msg2015. [Frederic Requin]
-* If VM_PARALLEL_BUILDS=1, use OPT_FAST and OPT_SLOW. [Frederic Requin]
+* Add --no-decoration to remove output comments, msg2015. [Frédéric Requin]
+* If VM_PARALLEL_BUILDS=1, use OPT_FAST and OPT_SLOW. [Frédéric Requin]
  Set VM_DEFAULT_RULES=0 for old behavior.
 * Add error on DPI functions > 32 bits. (#1898) [Elliot Mednick]
 * Improve Verilation performance on internal strings. (#1896) [Johan Bjork]
@@ -1478,7 +1910,7 @@ Verilator 3.884 2016-05-18
 * Fix --output-split of constructors. (#1035) [Johan Bjork]
 * Fix removal of empty packages, modules and cells. (#1034) [Johan Bjork]
 * Fix core dump on Arch Linux/GCC 6.1.1. (#1058) [Jannis Harder]
-* Fix $value$plusargs to string. (#1880) [Frederic Requin]
+* Fix $value$plusargs to string. (#1880) [Frédéric Requin]


 Verilator 3.882 2016-03-01
@@ -1557,13 +1989,13 @@ Verilator 3.876 2015-08-12

 **Minor:**

-* Add tracing_on, etc to vlt files. (#932) [Frederic Requin]
+* Add tracing_on, etc to vlt files. (#932) [Frédéric Requin]
 * Support extraction of enum bits. (#951) [Jonathon Donaldson]
 * Fix MinGW compiler error. (#927) (#929) [Hans Tichelaar]
 * Fix .c files to be treated as .cpp. (#930) [Jonathon Donaldson]
 * Fix string-to-int space conversion. (#931) [Fabrizio Ferrandi]
 * Fix dpi imports inside generates. [Michael Tresidder]
-* Fix rounding in trace $timescale. (#946) [Frederic Requin]
+* Fix rounding in trace $timescale. (#946) [Frédéric Requin]
 * Fix $fopen with SV string. (#947) [Sven Stucki]
 * Fix hashed error with typedef inside block. (#948) [Sven Stucki]
 * Fix makefile with --coverage. (#953) [Eivind Liland]
@@ -1803,7 +2235,7 @@ Verilator 3.852 2013-09-29

 * Support named function and task arguments. [Chris Randall]
 * Report SELRANGE warning for non-generate if. (#675) [Roland Kruse]
-* Fix ordering of $fgetc. (#1808) [Frederic Requin]
+* Fix ordering of $fgetc. (#1808) [Frédéric Requin]
 * Fix --output-split-cfunc to count internal functions. [Chris Randall]
 * Fix crash on 32-bit Ubuntu. (#670) [Mark Jackson Pulver]

@@ -1871,7 +2303,7 @@ Verilator 3.846 2013-03-09
 * Fix DETECTARRAY on packed structures. (#610) [Jeremy Bennett]
 * Fix LITENDIAN on unpacked structures. (#614) [Wai Sum Mong]
 * Fix 32-bit OS VPI scan issue. (#615)  [Jeremy Bennett, Rich Porter]
-* Fix opening a VerilatedVcdC file multiple times. (#1774) [Frederic Requin]
+* Fix opening a VerilatedVcdC file multiple times. (#1774) [Frédéric Requin]
 * Fix UNOPTFLAT circular array bounds crossing. (#630) [Jie Xu]


@@ -3183,7 +3615,7 @@ Verilator 3.460 2005-07-27 Beta
 * Fix false warning when a clock is constant.
 * Fix X/Z in decimal numbers. [Wim Michiels]
 * Fix genvar statements in non-named generate blocks.
-* Fix core dump when missing newline in `define. [David van der bokke]
+* Fix core dump when missing newline in `define. [David van der Bokke]


 Verilator 3.450 2005-07-12
@@ -4079,7 +4511,7 @@ Verilator 0.0 1994-07-08
 Copyright
 =========

-Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -7,7 +7,7 @@
 #
 #*****************************************************************************
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
@@ -15,13 +15,13 @@
 #
 #****************************************************************************/
 #
-# make all	to compile and build Verilator.
-# make install	to install it.
-# make TAGS	to update tags tables.
+# make all      to compile and build Verilator.
+# make install  to install it.
+# make TAGS     to update tags tables.
 #
 # make clean  or  make mostlyclean
 #      Delete all files from the current directory that are normally
-#      created by building the program.	 Don't delete the files that
+#      created by building the program.  Don't delete the files that
 #      record the configuration.  Also preserve files that could be made
 #      by building, but normally aren't because the distribution comes
 #      with them.
@@ -153,7 +153,7 @@ msg_test: all_nomsg
 	@echo

 .PHONY: test
-ifeq ($(CFG_WITH_LONGTESTS),yes)	# Local... Else don't burden users
+ifeq ($(CFG_WITH_LONGTESTS),yes)  # Local... Else don't burden users
 test: smoke-test test_regress
 # examples is part of test_regress's test_regress/t/t_a2_examples.pl
 # (because that allows it to run in parallel with other test_regress's)
@@ -173,6 +173,10 @@ smoke-test: all_nomsg
 test_regress: all_nomsg
 	$(MAKE) -C test_regress

+.PHONY: test-snap test-diff
+test-snap test-diff:
+	$(MAKE) -C test_regress $@
+
 examples: all_nomsg
 	for p in $(EXAMPLES) ; do \
 	  $(MAKE) -C $$p VERILATOR_ROOT=`pwd` || exit 10; \
@@ -424,17 +428,21 @@ YAPF_FLAGS = -i
 yapf:
 	$(YAPF) $(YAPF_FLAGS) $(PY_FILES)

-FLAKE8 = flake8
-FLAKE8_FLAGS = \
-	--extend-exclude=fastcov.py \
-	--ignore=E123,E129,E251,E402,E501,W503,W504,E701
-
 PYLINT = pylint
 PYLINT_FLAGS = --score=n --disable=R0801

+RUFF = ruff
+RUFF_FLAGS = check --ignore=E402,E501,E701
+
+# "make -k" so can see all tool result errors
 lint-py:
-	-$(FLAKE8) $(FLAKE8_FLAGS) $(PY_PROGRAMS)
-	-$(PYLINT) $(PYLINT_FLAGS) $(PY_PROGRAMS)
+	$(MAKE) -k lint-py-pylint lint-py-ruff
+
+lint-py-pylint:
+	$(PYLINT) $(PYLINT_FLAGS) $(PY_PROGRAMS)
+
+lint-py-ruff:
+	$(RUFF) $(RUFF_FLAGS) $(PY_PROGRAMS)

 format-pl-exec:
 	-chmod a+x test_regress/t/*.pl
@@ -451,8 +459,8 @@ install-msg:

 IN_WILD := ${srcdir}/*.in ${srcdir}/*/*.in

-# autoheader might not change config_build.h.in, so touch it
-${srcdir}/config_build.h: ${srcdir}/config_build.h.in configure
+# autoheader might not change config_package.h.in, so touch it
+${srcdir}/config_package.h: ${srcdir}/config_package.h.in configure
 	cd ${srcdir} && autoheader
 	touch $@
 Makefile: Makefile.in config.status $(IN_WILD)
@@ -462,7 +470,7 @@ config.status: configure
 	./config.status --recheck

 configure: configure.ac
-ifeq ($(CFG_WITH_CCWARN),yes)	# Local... Else don't burden users
+ifeq ($(CFG_WITH_CCWARN),yes)  # Local... Else don't burden users
 	autoconf --warnings=all
 else
 	autoconf
@@ -501,7 +509,7 @@ distclean maintainer-clean::
 TAGFILES=${srcdir}/*/*.cpp ${srcdir}/*/*.h ${srcdir}/*/*.in \
 	${srcdir}/*.in ${srcdir}/*.pod

-TAGS:	$(TAGFILES)
+TAGS: $(TAGFILES)
 	etags $(TAGFILES)

 .PHONY: doxygen
@@ -1,12 +1,16 @@
 .. Github doesn't render images unless absolute URL
 .. Do not know of a conditional tag, "only: github" nor "github display" works

+.. image:: https://img.shields.io/badge/Website-Verilator.org-181717.svg
+    :target: https://verilator.org
 .. image:: https://img.shields.io/badge/License-LGPL%20v3-blue.svg
-    :target: https://www.gnu.org/licenses/lgpl-3.0]
+    :target: https://www.gnu.org/licenses/lgpl-3.0
 .. image:: https://img.shields.io/badge/License-Artistic%202.0-0298c3.svg
    :target: https://opensource.org/licenses/Artistic-2.0
 .. image:: https://repology.org/badge/tiny-repos/verilator.svg?header=distro%20packages
    :target: https://repology.org/project/verilator/versions
+.. image:: https://img.shields.io/docker/pulls/verilator/verilator
+    :target: https://hub.docker.com/r/verilator/verilator
 .. image:: https://api.codacy.com/project/badge/Grade/fa78caa433c84a4ab9049c43e9debc6f
    :target: https://www.codacy.com/gh/verilator/verilator
 .. image:: https://codecov.io/gh/verilator/verilator/branch/master/graph/badge.svg
@@ -57,17 +61,16 @@ files, the "Verilated" code.

 These Verilated C++/SystemC files are then compiled by a C++ compiler
 (gcc/clang/MSVC++), optionally along with a user's own C++/SystemC wrapper
-file to instantiate the Verilated model. Executing the resulting executable
-performs the design simulation. Verilator also supports linking Verilated
-generated libraries, optionally encrypted, into other simulators.
+file, to instantiate the Verilated model. Executing the resulting
+executable performs the design simulation. Verilator also supports linking
+Verilated generated libraries, optionally encrypted, into other simulators.

 Verilator may not be the best choice if you are expecting a full-featured
 replacement for a closed-source Verilog simulator, needs SDF annotation,
 mixed-signal simulation, or are doing a quick class project (we recommend
 `Icarus Verilog`_ for classwork.)  However, if you are looking for a path
 to migrate SystemVerilog to C++/SystemC, or want high-speed simulation of
-synthesizable designs containing limited verification constructs, Verilator
-is the tool for you.
+designs, Verilator is the tool for you.


 Performance
@@ -83,11 +86,11 @@ as `Icarus Verilog`_. Another 2-10x speedup might be gained from
 multithreading (yielding 200-1000x total over interpreted simulators).

 Verilator has typically similar or better performance versus the
-closed-source Verilog simulators (Carbon Design Systems Carbonator,
+closed-source Verilog simulators (e.g., Carbon Design Systems Carbonator,
 Modelsim/Questa, Cadence Incisive/NC-Verilog, Synopsys VCS, VTOC, and
 Pragmatic CVer/CVC). But, Verilator is open-sourced, so you can spend on
 computes rather than licenses. Thus, Verilator gives you the best
-cycles/dollar.
+simulation cycles/dollar.


 Installation & Documentation
@@ -139,7 +142,7 @@ Related Projects
 Open License
 ============

-Verilator is Copyright 2003-2023 by Wilson Snyder. (Report bugs to
+Verilator is Copyright 2003-2024 by Wilson Snyder. (Report bugs to
 `Verilator Issues <https://verilator.org/issues>`_.)

 Verilator is free software; you can redistribute it and/or modify it under
@@ -147,7 +150,7 @@ the terms of either the GNU Lesser General Public License Version 3 or the
 Perl Artistic License Version 2.0. See the documentation for more details.

 .. _CHIPS Alliance: https://chipsalliance.org
-.. _Icarus Verilog: http://iverilog.icarus.com
+.. _Icarus Verilog: https://steveicarus.github.io/iverilog
 .. _Linux Foundation: https://www.linuxfoundation.org
 .. |Logo| image:: https://www.veripool.org/img/verilator_256_200_min.png
 .. |verilator multithreaded performance| image:: https://www.veripool.org/img/verilator_multithreaded_performance_bg-min.png
@@ -1,7 +1,7 @@
 #!/usr/bin/env perl
 ######################################################################
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
@@ -362,7 +362,7 @@ detailed descriptions of these arguments.
     +incdir+<dir>              Directory to search for includes
    --inline-mult <value>       Tune module inlining
    --instr-count-dpi <value>   Assumed dynamic instruction count of DPI imports
-     -j <jobs>                  Parallelism for --build (alias to --build-jobs)
+     -j <jobs>                  Parallelism for --build-jobs/--verilate-jobs
    --l2-name <value>           Verilog scope name of the top module
    --language <lang>           Default language standard to parse
     -LDFLAGS <flags>           Linker pre-object arguments for makefile
@@ -372,6 +372,7 @@ detailed descriptions of these arguments.
    --make <build-tool>         Generate scripts for specified build tool
     -MAKEFLAGS <flags>         Arguments to pass to make during --build
    --main                      Generate C++ main() file
+    --main-top-name             Specify top name passed to Verilated model in generated C++ main
    --max-num-width <value>     Maximum number width (default: 64K)
    --Mdir <directory>          Name of output object directory
    --MMD                       Create .d dependency files
@@ -404,6 +405,7 @@ detailed descriptions of these arguments.
    --protect-key <key>         Key for symbol protection
    --protect-lib <name>        Create a DPI protected library
    --public                    Debugging; see docs
+    --public-params             Mark all parameters as public_flat
    --public-flat-rw            Mark all variables, etc as public_flat_rw
     -pvalue+<name>=<value>     Overwrite toplevel parameter
    --quiet-exit                Don't print the command on failure
@@ -416,6 +418,7 @@ detailed descriptions of these arguments.
    --no-skip-identical         Disable skipping identical output
    --stats                     Create statistics file
    --stats-vars                Provide statistics on variables
+    --no-std                    Prevent parsing standard library
    --structs-packed            Convert all unpacked structures to packed structures
     -sv                        Enable SystemVerilog parsing
     +systemverilogext+<ext>    Synonym for +1800-2017ext+<ext>
@@ -437,6 +440,7 @@ detailed descriptions of these arguments.
    --trace-params              Enable tracing of parameters
    --trace-structs             Enable tracing structure names
    --trace-threads <threads>   Enable FST waveform creation on separate threads
+    --no-trace-top              Do not emit traces for signals in the top module generated by verilator
    --trace-underscore          Enable tracing of _signals
     -U<var>                    Undefine preprocessor define
    --no-unlimited-stack        Don't disable stack size limit
@@ -445,6 +449,7 @@ detailed descriptions of these arguments.
    --unused-regexp <regexp>    Tune UNUSED lint signals
     -V                         Verbose version and config
     -v <filename>              Verilog library
+    --verilate-jobs             Job threads for Verilation stage
    --no-verilate               Skip Verilation and just compile previously Verilated code
     +verilog1995ext+<ext>      Synonym for +1364-1995ext+<ext>
     +verilog2001ext+<ext>      Synonym for +1364-2001ext+<ext>
@@ -496,7 +501,7 @@ description of these arguments.

 The latest version is available from L<https://verilator.org>.

-Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 redistribute it and/or modify the Verilator internals under the terms of
 either the GNU Lesser General Public License Version 3 or the Perl Artistic
 License Version 2.0.
@@ -17,7 +17,7 @@ parser = argparse.ArgumentParser(
    For documentation see
    https://verilator.org/guide/latest/exe_verilator_ccache_report.html""",
    epilog=
-    """Copyright 2002-2023 by Wilson Snyder. This program is free software; you
+    """Copyright 2002-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -1,7 +1,7 @@
 #!/usr/bin/env perl
 ######################################################################
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
@@ -169,6 +169,7 @@ L<https://verilator.org/guide/latest/exe_verilator_coverage.html>.
    --annotate <output_dir>       Directory name for source annotation.
    --annotate-all                All files should be shown.
    --annotate-min <count>        Minimum occurrence count for uncovered.
+    --annotate-points             Annotates info from each coverage point.
    --help                        Displays this message and version and exits.
    --rank                        Compute relative importance of tests.
    --unlink                      With --write, unlink all inputs
@@ -187,7 +188,7 @@ L<https://verilator.org/guide/latest/exe_verilator_coverage.html>.

 The latest version is available from L<https://verilator.org>.

-Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify the Verilator internals under the terms
 of either the GNU Lesser General Public License Version 3 or the Perl
 Artistic License Version 2.0.
@@ -42,8 +42,8 @@ def diff_dir(a, b):
            continue
        a = files[base]['a']
        b = files[base]['b']
-        print("=" * 70)
-        print("= %s <-> %s" % (a, b))
+        print("=" * 70, flush=True)
+        print("= %s <-> %s" % (a, b), flush=True)
        diff_file(a, b)
        anyfile = True
    if not anyfile:
@@ -110,7 +110,7 @@ Verilator_difftree is used for debugging Verilator tree output files.
 It performs a diff between two files, or all files common between two
 directories, ignoring irrelevant pointer differences.""",
    epilog=
-    """Copyright 2005-2023 by Wilson Snyder. This program is free software; you
+    """Copyright 2005-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -1,33 +1,32 @@
 #!/usr/bin/env python3
-# pylint: disable=C0103,C0114,C0116,C0209,C0301,R0914,R0912,R0915,W0511,eval-used
+# pylint: disable=C0103,C0114,C0116,C0209,C0301,R0914,R0912,R0915,W0511,W0603,eval-used
 ######################################################################

 import argparse
+import bisect
 import collections
 import math
 import re
 import statistics
 # from pprint import pprint

-Threads = collections.defaultdict(lambda: collections.defaultdict(lambda: {}))
-Mtasks = collections.defaultdict(lambda: {})
-Evals = collections.defaultdict(lambda: {})
-EvalLoops = collections.defaultdict(lambda: {})
+Sections = []
+LongestVcdStrValueLength = 0
+Threads = collections.defaultdict(lambda: [])  # List of records per thread id
+Mtasks = collections.defaultdict(lambda: {'elapsed': 0, 'end': 0})
+Cpus = collections.defaultdict(lambda: {'mtask_time': 0})
 Global = {
    'args': {},
    'cpuinfo': collections.defaultdict(lambda: {}),
-    'rdtsc_cycle_time': 0,
    'stats': {}
 }
+ElapsedTime = None  # total elapsed time
+ExecGraphTime = 0  # total elapsed time excuting an exec graph
+ExecGraphIntervals = []  # list of (start, end) pairs

 ######################################################################


-def process(filename):
-    read_data(filename)
-    report()
-
-
 def read_data(filename):
    with open(filename, "r", encoding="utf8") as fh:
        re_thread = re.compile(r'^VLPROFTHREAD (\d+)$')
@@ -39,14 +38,17 @@ def read_data(filename):
        re_arg1 = re.compile(r'VLPROF arg\s+(\S+)\+([0-9.]*)\s*')
        re_arg2 = re.compile(r'VLPROF arg\s+(\S+)\s+([0-9.]*)\s*$')
        re_stat = re.compile(r'VLPROF stat\s+(\S+)\s+([0-9.]+)')
-        re_time = re.compile(r'rdtsc time = (\d+) ticks')
        re_proc_cpu = re.compile(r'VLPROFPROC processor\s*:\s*(\d+)\s*$')
        re_proc_dat = re.compile(r'VLPROFPROC ([a-z_ ]+)\s*:\s*(.*)$')
        cpu = None
        thread = None
+        execGraphStart = None

-        lastEvalBeginTick = None
-        lastEvalLoopBeginTick = None
+        global LongestVcdStrValueLength
+        global ExecGraphTime
+
+        SectionStack = []
+        mTaskThread = {}

        for line in fh:
            recordMatch = re_record.match(line)
@@ -54,29 +56,31 @@ def read_data(filename):
                kind, tick, payload = recordMatch.groups()
                tick = int(tick)
                payload = payload.strip()
-                if kind == "EVAL_BEGIN":
-                    Evals[tick]['start'] = tick
-                    lastEvalBeginTick = tick
-                elif kind == "EVAL_END":
-                    Evals[lastEvalBeginTick]['end'] = tick
-                    lastEvalBeginTick = None
-                elif kind == "EVAL_LOOP_BEGIN":
-                    EvalLoops[tick]['start'] = tick
-                    lastEvalLoopBeginTick = tick
-                elif kind == "EVAL_LOOP_END":
-                    EvalLoops[lastEvalLoopBeginTick]['end'] = tick
-                    lastEvalLoopBeginTick = None
+                if kind == "SECTION_PUSH":
+                    LongestVcdStrValueLength = max(LongestVcdStrValueLength,
+                                                   len(payload))
+                    SectionStack.append(payload)
+                    Sections.append((tick, tuple(SectionStack)))
+                elif kind == "SECTION_POP":
+                    assert SectionStack, "SECTION_POP without SECTION_PUSH"
+                    SectionStack.pop()
+                    Sections.append((tick, tuple(SectionStack)))
                elif kind == "MTASK_BEGIN":
                    mtask, predict_start, ecpu = re_payload_mtaskBegin.match(
                        payload).groups()
                    mtask = int(mtask)
                    predict_start = int(predict_start)
                    ecpu = int(ecpu)
-                    Threads[thread][tick]['mtask'] = mtask
-                    Threads[thread][tick]['predict_start'] = predict_start
-                    Threads[thread][tick]['cpu'] = ecpu
-                    if 'elapsed' not in Mtasks[mtask]:
-                        Mtasks[mtask] = {'end': 0, 'elapsed': 0}
+                    mTaskThread[mtask] = thread
+                    records = Threads[thread]
+                    assert not records or records[-1]['start'] <= records[-1][
+                        'end'] <= tick
+                    records.append({
+                        'start': tick,
+                        'mtask': mtask,
+                        'predict_start': predict_start,
+                        'cpu': ecpu
+                    })
                    Mtasks[mtask]['begin'] = tick
                    Mtasks[mtask]['thread'] = thread
                    Mtasks[mtask]['predict_start'] = predict_start
@@ -86,11 +90,18 @@ def read_data(filename):
                    mtask = int(mtask)
                    predict_cost = int(predict_cost)
                    begin = Mtasks[mtask]['begin']
-                    Threads[thread][begin]['end'] = tick
-                    Threads[thread][begin]['predict_cost'] = predict_cost
+                    record = Threads[mTaskThread[mtask]][-1]
+                    record['end'] = tick
+                    record['predict_cost'] = predict_cost
                    Mtasks[mtask]['elapsed'] += tick - begin
                    Mtasks[mtask]['predict_cost'] = predict_cost
                    Mtasks[mtask]['end'] = max(Mtasks[mtask]['end'], tick)
+                elif kind == "EXEC_GRAPH_BEGIN":
+                    execGraphStart = tick
+                elif kind == "EXEC_GRAPH_END":
+                    ExecGraphTime += tick - execGraphStart
+                    ExecGraphIntervals.append((execGraphStart, tick))
+                    execGraphStart = None
                elif Args.debug:
                    print("-Unknown execution trace record: %s" % line)
            elif re_thread.match(line):
@@ -109,7 +120,7 @@ def read_data(filename):
            elif re_proc_cpu.match(line):
                match = re_proc_cpu.match(line)
                cpu = int(match.group(1))
-            elif cpu and re_proc_dat.match(line):
+            elif cpu is not None and re_proc_dat.match(line):
                match = re_proc_dat.match(line)
                term = match.group(1)
                value = match.group(2)
@@ -121,11 +132,6 @@ def read_data(filename):
                pass
            elif Args.debug:
                print("-Unk: %s" % line)
-            # TODO -- this is parsing text printed by a client.
-            # Really, verilator proper should generate this
-            # if it's useful...
-            if re_time.match(line):
-                Global['rdtsc_cycle_time'] = re_time.group(1)


 def re_match_result(regexp, line, result_to):
@@ -144,125 +150,33 @@ def report():
        plus = "+" if re.match(r'^\+', arg) else " "
        print("  %s%s%s" % (arg, plus, Global['args'][arg]))

+    for records in Threads.values():
+        for record in records:
+            cpu = record['cpu']
+            elapsed = record['end'] - record['start']
+            Cpus[cpu]['mtask_time'] += elapsed
+
+    global ElapsedTime
+    ElapsedTime = int(Global['stats']['ticks'])
    nthreads = int(Global['stats']['threads'])
-    Global['cpus'] = {}
-    for thread in Threads:
-        # Make potentially multiple characters per column
-        for start in Threads[thread]:
-            if not Threads[thread][start]:
-                continue
-            cpu = Threads[thread][start]['cpu']
-            elapsed = Threads[thread][start]['end'] - start
-            if cpu not in Global['cpus']:
-                Global['cpus'][cpu] = {'cpu_time': 0}
-            Global['cpus'][cpu]['cpu_time'] += elapsed
+    ncpus = max(len(Cpus), 1)

-    measured_mt_mtask_time = 0
-    predict_mt_mtask_time = 0
-    long_mtask_time = 0
-    measured_last_end = 0
-    predict_last_end = 0
-    for mtask in Mtasks:
-        measured_mt_mtask_time += Mtasks[mtask]['elapsed']
-        predict_mt_mtask_time += Mtasks[mtask]['predict_cost']
-        measured_last_end = max(measured_last_end, Mtasks[mtask]['end'])
-        predict_last_end = max(
-            predict_last_end,
-            Mtasks[mtask]['predict_start'] + Mtasks[mtask]['predict_cost'])
-        long_mtask_time = max(long_mtask_time, Mtasks[mtask]['elapsed'])
-    Global['measured_last_end'] = measured_last_end
-    Global['predict_last_end'] = predict_last_end
-
-    # If we know cycle time in the same (rdtsc) units,
-    # this will give us an actual utilization number,
-    # (how effectively we keep the cores busy.)
-    #
-    # It also gives us a number we can compare against
-    # serial mode, to estimate the overhead of data sharing,
-    # which will show up in the total elapsed time. (Overhead
-    # of synchronization and scheduling should not.)
-    print("\nAnalysis:")
-    print("  Total threads             = %d" % nthreads)
-    print("  Total mtasks              = %d" % len(Mtasks))
-    ncpus = max(len(Global['cpus']), 1)
-    print("  Total cpus used           = %d" % ncpus)
-    print("  Total yields              = %d" %
-          int(Global['stats'].get('yields', 0)))
-    print("  Total evals               = %d" % len(Evals))
-    print("  Total eval loops          = %d" % len(EvalLoops))
-    if Mtasks:
-        print("  Total eval time           = %d rdtsc ticks" %
-              Global['measured_last_end'])
-        print("  Longest mtask time        = %d rdtsc ticks" % long_mtask_time)
-        print("  All-thread mtask time     = %d rdtsc ticks" %
-              measured_mt_mtask_time)
-        long_efficiency = long_mtask_time / (Global.get(
-            'measured_last_end', 1) or 1)
-        print("  Longest-thread efficiency = %0.1f%%" %
-              (long_efficiency * 100.0))
-        mt_efficiency = measured_mt_mtask_time / (
-            Global.get('measured_last_end', 1) * nthreads or 1)
-        print("  All-thread efficiency     = %0.1f%%" %
-              (mt_efficiency * 100.0))
-        print("  All-thread speedup        = %0.1f" %
-              (mt_efficiency * nthreads))
-        if Global['rdtsc_cycle_time'] > 0:
-            ut = measured_mt_mtask_time / Global['rdtsc_cycle_time']
-            print("tot_mtask_cpu=" + measured_mt_mtask_time + " cyc=" +
-                  Global['rdtsc_cycle_time'] + " ut=" + ut)
-
-        predict_mt_efficiency = predict_mt_mtask_time / (
-            Global.get('predict_last_end', 1) * nthreads or 1)
-        print("\nPrediction (what Verilator used for scheduling):")
-        print("  All-thread efficiency     = %0.1f%%" %
-              (predict_mt_efficiency * 100.0))
-        print("  All-thread speedup        = %0.1f" %
-              (predict_mt_efficiency * nthreads))
-
-        p2e_ratios = []
-        min_p2e = 1000000
-        min_mtask = None
-        max_p2e = -1000000
-        max_mtask = None
-
-        for mtask in sorted(Mtasks.keys()):
-            if Mtasks[mtask]['elapsed'] > 0:
-                if Mtasks[mtask]['predict_cost'] == 0:
-                    Mtasks[mtask]['predict_cost'] = 1  # don't log(0) below
-                p2e_ratio = math.log(Mtasks[mtask]['predict_cost'] /
-                                     Mtasks[mtask]['elapsed'])
-                p2e_ratios.append(p2e_ratio)
-
-                if p2e_ratio > max_p2e:
-                    max_p2e = p2e_ratio
-                    max_mtask = mtask
-                if p2e_ratio < min_p2e:
-                    min_p2e = p2e_ratio
-                    min_mtask = mtask
-
-        print("\nMTask statistics:")
-        print("  min log(p2e) = %0.3f" % min_p2e, end="")
-        print("  from mtask %d (predict %d," %
-              (min_mtask, Mtasks[min_mtask]['predict_cost']),
-              end="")
-        print(" elapsed %d)" % Mtasks[min_mtask]['elapsed'])
-        print("  max log(p2e) = %0.3f" % max_p2e, end="")
-        print("  from mtask %d (predict %d," %
-              (max_mtask, Mtasks[max_mtask]['predict_cost']),
-              end="")
-        print(" elapsed %d)" % Mtasks[max_mtask]['elapsed'])
-
-        stddev = statistics.pstdev(p2e_ratios)
-        mean = statistics.mean(p2e_ratios)
-        print("  mean = %0.3f" % mean)
-        print("  stddev = %0.3f" % stddev)
-        print("  e ^ stddev = %0.3f" % math.exp(stddev))
+    print("\nSummary:")
+    print("  Total elapsed time = {} rdtsc ticks".format(ElapsedTime))
+    print("  Parallelized code  = {:.2%} of elapsed time".format(
+        ExecGraphTime / ElapsedTime))
+    print("  Total threads      = %d" % nthreads)
+    print("  Total CPUs used    = %d" % ncpus)
+    print("  Total mtasks       = %d" % len(Mtasks))
+    print("  Total yields       = %d" % int(Global['stats'].get('yields', 0)))

+    report_mtasks()
    report_cpus()
+    report_sections()

    if nthreads > ncpus:
        print()
-        print("%%Warning: There were fewer CPUs (%d) then threads (%d)." %
+        print("%%Warning: There were fewer CPUs (%d) than threads (%d)." %
              (ncpus, nthreads))
        print("        : See docs on use of numactl.")
    else:
@@ -279,32 +193,130 @@ def report():
    print()


+def report_mtasks():
+    if not Mtasks:
+        return
+
+    nthreads = int(Global['stats']['threads'])
+
+    # If we know cycle time in the same (rdtsc) units,
+    # this will give us an actual utilization number,
+    # (how effectively we keep the cores busy.)
+    #
+    # It also gives us a number we can compare against
+    # serial mode, to estimate the overhead of data sharing,
+    # which will show up in the total elapsed time. (Overhead
+    # of synchronization and scheduling should not.)
+    total_mtask_time = 0
+    thread_mtask_time = collections.defaultdict(lambda: 0)
+    long_mtask_time = 0
+    long_mtask = None
+    predict_mtask_time = 0
+    predict_elapsed = 0
+    for mtaskId in Mtasks:
+        record = Mtasks[mtaskId]
+        predict_mtask_time += record['predict_cost']
+        total_mtask_time += record['elapsed']
+        thread_mtask_time[record['thread']] += record['elapsed']
+        predict_end = record['predict_start'] + record['predict_cost']
+        predict_elapsed = max(predict_elapsed, predict_end)
+        if record['elapsed'] > long_mtask_time:
+            long_mtask_time = record['elapsed']
+            long_mtask = mtaskId
+    Global['predict_last_end'] = predict_elapsed
+
+    serialTime = ElapsedTime - ExecGraphTime
+
+    def subReport(elapsed, work):
+        print("  Thread utilization = {:7.2%}".format(work /
+                                                      (elapsed * nthreads)))
+        print("  Speedup            = {:6.3}x".format(work / elapsed))
+
+    print("\nParallelized code, measured:")
+    subReport(ExecGraphTime, total_mtask_time)
+
+    print("\nParallelized code, predicted during static scheduling:")
+    subReport(predict_elapsed, predict_mtask_time)
+
+    print("\nAll code, measured:")
+    subReport(ElapsedTime, serialTime + total_mtask_time)
+
+    print("\nAll code, measured, scaled by predicted speedup:")
+    expectedParallelSpeedup = predict_mtask_time / predict_elapsed
+    scaledElapsed = serialTime + total_mtask_time / expectedParallelSpeedup
+    subReport(scaledElapsed, serialTime + total_mtask_time)
+
+    p2e_ratios = []
+    min_p2e = 1000000
+    min_mtask = None
+    max_p2e = -1000000
+    max_mtask = None
+
+    for mtask in sorted(Mtasks.keys()):
+        if Mtasks[mtask]['elapsed'] > 0:
+            if Mtasks[mtask]['predict_cost'] == 0:
+                Mtasks[mtask]['predict_cost'] = 1  # don't log(0) below
+            p2e_ratio = math.log(Mtasks[mtask]['predict_cost'] /
+                                 Mtasks[mtask]['elapsed'])
+            p2e_ratios.append(p2e_ratio)
+
+            if p2e_ratio > max_p2e:
+                max_p2e = p2e_ratio
+                max_mtask = mtask
+            if p2e_ratio < min_p2e:
+                min_p2e = p2e_ratio
+                min_mtask = mtask
+
+    print("\nMTask statistics:")
+    print("  Longest mtask id = {}".format(long_mtask))
+    print("  Longest mtask time = {:.2%} of time elapsed in parallelized code".
+          format(long_mtask_time / ExecGraphTime))
+    print("  min log(p2e) = %0.3f" % min_p2e, end="")
+
+    print("  from mtask %d (predict %d," %
+          (min_mtask, Mtasks[min_mtask]['predict_cost']),
+          end="")
+    print(" elapsed %d)" % Mtasks[min_mtask]['elapsed'])
+    print("  max log(p2e) = %0.3f" % max_p2e, end="")
+    print("  from mtask %d (predict %d," %
+          (max_mtask, Mtasks[max_mtask]['predict_cost']),
+          end="")
+    print(" elapsed %d)" % Mtasks[max_mtask]['elapsed'])
+
+    stddev = statistics.pstdev(p2e_ratios)
+    mean = statistics.mean(p2e_ratios)
+    print("  mean = %0.3f" % mean)
+    print("  stddev = %0.3f" % stddev)
+    print("  e ^ stddev = %0.3f" % math.exp(stddev))
+
+
 def report_cpus():
-    print("\nCPUs:")
+    print("\nCPU info:")

    Global['cpu_sockets'] = collections.defaultdict(lambda: 0)
    Global['cpu_socket_cores'] = collections.defaultdict(lambda: 0)

-    for cpu in sorted(Global['cpus'].keys()):
-        print("  cpu %d: " % cpu, end='')
-        print("cpu_time=%d" % Global['cpus'][cpu]['cpu_time'], end='')
-
-        socket = None
+    print("   Id | Time spent executing MTask | Socket | Core | Model")
+    print("      | % of elapsed ticks / ticks |        |      |")
+    print("  ====|============================|========|======|======")
+    for cpu in sorted(Cpus):
+        socket = ""
+        core = ""
+        model = ""
        if cpu in Global['cpuinfo']:
            cpuinfo = Global['cpuinfo'][cpu]
            if 'physical_id' in cpuinfo and 'core_id' in cpuinfo:
-                socket = int(cpuinfo['physical_id'])
+                socket = cpuinfo['physical_id']
                Global['cpu_sockets'][socket] += 1
-                print(" socket=%d" % socket, end='')
-
-                core = int(cpuinfo['core_id'])
-                Global['cpu_socket_cores'][str(socket) + "__" + str(core)] += 1
-                print(" core=%d" % core, end='')
+                core = cpuinfo['core_id']
+                Global['cpu_socket_cores'][socket + "__" + core] += 1

            if 'model_name' in cpuinfo:
                model = cpuinfo['model_name']
-                print("  %s" % model, end='')
-        print()
+
+        print("  {:3d} | {:7.2%} / {:16d} | {:>6s} | {:>4s} | {}".format(
+            cpu, Cpus[cpu]['mtask_time'] / ElapsedTime,
+            Cpus[cpu]['mtask_time'], socket, core, model))

    if len(Global['cpu_sockets']) > 1:
        Global['cpu_sockets_warning'] = True
@@ -313,25 +325,68 @@ def report_cpus():
                Global['cpu_socket_cores_warning'] = True


+def report_sections():
+    if not Sections:
+        return
+    print("\nSection profile:")
+
+    totalTime = collections.defaultdict(lambda: 0)
+    selfTime = collections.defaultdict(lambda: 0)
+
+    sectionTree = [0, {}, 1]  # [selfTime, childTrees, numberOfTimesEntered]
+    prevTime = 0
+    prevStack = ()
+    for time, stack in Sections:
+        if len(stack) > len(prevStack):
+            scope = sectionTree
+            for item in stack:
+                scope = scope[1].setdefault(item, [0, {}, 0])
+            scope[2] += 1
+        dt = time - prevTime
+        scope = sectionTree
+        for item in prevStack:
+            scope = scope[1].setdefault(item, [0, {}, 0])
+        scope[0] += dt
+
+        if prevStack:
+            for name in prevStack:
+                totalTime[name] += dt
+            selfTime[prevStack[-1]] += dt
+        prevTime = time
+        prevStack = stack
+
+    def treeSum(tree):
+        n = tree[0]
+        for subTree in tree[1].values():
+            n += treeSum(subTree)
+        return n
+
+    # Make sure the tree sums to the elapsed time
+    sectionTree[0] += ElapsedTime - treeSum(sectionTree)
+
+    def printTree(prefix, name, entries, tree):
+        print("  {:7.2%} | {:7.2%} | {:8} | {:10.2f} | {}".format(
+            treeSum(tree) / ElapsedTime, tree[0] / ElapsedTime, tree[2],
+            tree[2] / entries, prefix + name))
+        for k in sorted(tree[1], key=lambda _: -treeSum(tree[1][_])):
+            printTree(prefix + "  ", k, tree[2], tree[1][k])
+
+    print(" Total    | Self    | Total    | Relative   | Section")
+    print(" time     | time    | entries  | entries    |  name  ")
+    print("==========|=========|==========|============|========")
+    printTree("", "*TOTAL*", 1, sectionTree)
+
+
 ######################################################################


 def write_vcd(filename):
    print("Writing %s" % filename)
    with open(filename, "w", encoding="utf8") as fh:
-        vcd = {
-            'values':
-            collections.defaultdict(lambda: {}),  # {<time>}{<code>} = value
-            'sigs': {
-                'predicted_threads': {},
-                'measured_threads': {},
-                'cpus': {},
-                'evals': {},
-                'mtasks': {},
-                'Stats': {}
-            }  # {<module>}{<sig}} = code
-        }
-        code = 0
+        # dict of dicts of hierarchy elements/signal name  -> (code, width)
+        scopeSigs = {}
+        # map from time -> code -> value
+        values = collections.defaultdict(lambda: {})

        parallelism = {
            'measured': collections.defaultdict(lambda: 0),
@@ -340,132 +395,139 @@ def write_vcd(filename):
        parallelism['measured'][0] = 0
        parallelism['predicted'][0] = 0

-        # Measured graph
-        for thread in sorted(Threads.keys()):
-            sig = "thread%d_mtask" % thread
-            if sig not in vcd['sigs']['measured_threads']:
-                vcd['sigs']['measured_threads'][sig] = code
-                code += 1
-            mcode = vcd['sigs']['measured_threads'][sig]
+        nextFreeCode = 0

-            for start in sorted(Threads[thread]):
-                mtask = Threads[thread][start]['mtask']
-                end = Threads[thread][start]['end']
-                cpu = Threads[thread][start]['cpu']
-                vcd['values'][start][mcode] = mtask
-                vcd['values'][end][mcode] = None
+        def getCode(width, *names):
+            nonlocal nextFreeCode
+            scope = scopeSigs
+            *path, name = names
+            for item in path:
+                scope = scope.setdefault(item, {})
+            code, oldWidth = scope.setdefault(name, (nextFreeCode, width))
+            assert oldWidth == width
+            if code == nextFreeCode:
+                nextFreeCode += 1
+            return code
+
+        def addValue(code, time, val):
+            if isinstance(val, str):
+                buf = ""
+                for c in val:
+                    buf += bin(ord(c))[2:].rjust(8, "0")
+                val = buf.ljust(LongestVcdStrValueLength * 8, "0")
+            values[time][code] = val
+
+        # Measured graph
+        for thread in sorted(Threads):
+            mcode = getCode(32, 'measured', 't%d_mtask' % thread)
+            for record in Threads[thread]:
+                start = record['start']
+                mtask = record['mtask']
+                end = record['end']
+                cpu = record['cpu']
+                addValue(mcode, start, mtask)
+                addValue(mcode, end, None)
                parallelism['measured'][start] += 1
                parallelism['measured'][end] -= 1

-                sig = "cpu%d_thread" % cpu
-                if sig not in vcd['sigs']['cpus']:
-                    vcd['sigs']['cpus'][sig] = code
-                    code += 1
-                ccode = vcd['sigs']['cpus'][sig]
-                vcd['values'][start][ccode] = thread
-                vcd['values'][end][ccode] = None
+                code = getCode(32, 'measured', 'cpu%d_thread' % cpu)
+                addValue(code, start, thread)
+                addValue(code, end, None)

-                sig = "mtask%d_cpu" % mtask
-                if sig not in vcd['sigs']['mtasks']:
-                    vcd['sigs']['mtasks'][sig] = code
-                    code += 1
-                ccode = vcd['sigs']['mtasks'][sig]
-                vcd['values'][start][ccode] = cpu
-                vcd['values'][end][ccode] = None
+                code = getCode(32, 'measured', 'cpu%d_mtask' % cpu)
+                addValue(code, start, mtask)
+                addValue(code, end, None)

-        # Eval graph
-        vcd['sigs']['evals']["eval"] = code
-        elcode = code
-        code += 1
-        n = 0
-        for eval_start in Evals:
-            eval_end = Evals[eval_start]['end']
-            n += 1
-            vcd['values'][eval_start][elcode] = n
-            vcd['values'][eval_end][elcode] = None
+        tStart = sorted(_['start'] for records in Threads.values()
+                        for _ in records)
+        tEnd = sorted(_['end'] for records in Threads.values()
+                      for _ in records)

-        # Eval_loop graph
-        vcd['sigs']['evals']["eval_loop"] = code
-        elcode = code
-        code += 1
-        n = 0
-        for eval_start in EvalLoops:
-            eval_end = EvalLoops[eval_start]['end']
-            n += 1
-            vcd['values'][eval_start][elcode] = n
-            vcd['values'][eval_end][elcode] = None
+        # Predicted graph
+        for start, end in ExecGraphIntervals:
+            # Find the earliest MTask start after the start point, and the
+            # latest MTask end before the end point, so we can scale to the
+            # same range
+            start = tStart[bisect.bisect_left(tStart, start)]
+            end = tEnd[bisect.bisect_right(tEnd, end) - 1]
+            # Compute scale so predicted graph is of same width as interval
+            measured_scaling = (end - start) / Global['predict_last_end']
+            # Predict mtasks that fill the time the execution occupied
+            for mtask in Mtasks:
+                thread = Mtasks[mtask]['thread']
+                pred_scaled_start = start + int(
+                    Mtasks[mtask]['predict_start'] * measured_scaling)
+                pred_scaled_end = start + int(
+                    (Mtasks[mtask]['predict_start'] +
+                     Mtasks[mtask]['predict_cost']) * measured_scaling)
+                if pred_scaled_start == pred_scaled_end:
+                    continue

-        if Mtasks:
-            # Predicted graph
-            for eval_start in EvalLoops:
-                eval_end = EvalLoops[eval_start]['end']
-                # Compute scale so predicted graph is of same width as eval
-                measured_scaling = (eval_end -
-                                    eval_start) / Global['predict_last_end']
-                # Predict mtasks that fill the time the eval occupied
-                for mtask in Mtasks:
-                    thread = Mtasks[mtask]['thread']
-                    pred_scaled_start = eval_start + int(
-                        Mtasks[mtask]['predict_start'] * measured_scaling)
-                    pred_scaled_end = eval_start + int(
-                        (Mtasks[mtask]['predict_start'] +
-                         Mtasks[mtask]['predict_cost']) * measured_scaling)
-                    if pred_scaled_start == pred_scaled_end:
-                        continue
+                mcode = getCode(32, 'predicted', 't%d_mtask' % thread)
+                addValue(mcode, pred_scaled_start, mtask)
+                addValue(mcode, pred_scaled_end, None)

-                    sig = "predicted_thread%d_mtask" % thread
-                    if sig not in vcd['sigs']['predicted_threads']:
-                        vcd['sigs']['predicted_threads'][sig] = code
-                        code += 1
-                    mcode = vcd['sigs']['predicted_threads'][sig]
+                parallelism['predicted'][pred_scaled_start] += 1
+                parallelism['predicted'][pred_scaled_end] -= 1

-                    vcd['values'][pred_scaled_start][mcode] = mtask
-                    vcd['values'][pred_scaled_end][mcode] = None
+        # Parallelism graph
+        for measpred in ('measured', 'predicted'):
+            pcode = getCode(32, 'stats', '%s_parallelism' % measpred)
+            value = 0
+            for time in sorted(parallelism[measpred].keys()):
+                value += parallelism[measpred][time]
+                addValue(pcode, time, value)

-                    parallelism['predicted'][pred_scaled_start] += 1
-                    parallelism['predicted'][pred_scaled_end] -= 1
-
-            # Parallelism graph
-            for measpred in ('measured', 'predicted'):
-                vcd['sigs']['Stats']["%s_parallelism" % measpred] = code
-                pcode = code
-                code += 1
-                value = 0
-                for time in sorted(parallelism[measpred].keys()):
-                    value += parallelism[measpred][time]
-                    vcd['values'][time][pcode] = value
+        # Section graph
+        if Sections:
+            scode = getCode(LongestVcdStrValueLength * 8, "section", "trace")
+            dcode = getCode(32, "section", "depth")
+            for time, stack in Sections:
+                addValue(scode, time, stack[-1] if stack else None)
+                addValue(dcode, time, len(stack))

        # Create output file
        fh.write("$version Generated by verilator_gantt $end\n")
        fh.write("$timescale 1ns $end\n")
        fh.write("\n")

-        all_codes = {}
+        all_codes = set()
+
+        def writeScope(scope):
+            assert isinstance(scope, dict)
+            for key in sorted(scope):
+                val = scope[key]
+                if isinstance(val, dict):
+                    fh.write("  $scope module %s $end\n" % key)
+                    writeScope(val)
+                    fh.write("  $upscope $end\n")
+                else:
+                    (code, width) = val
+                    all_codes.add(code)
+                    fh.write("   $var wire %d v%x %s [%d:0] $end\n" %
+                             (width, code, key, width - 1))
+
        fh.write(" $scope module gantt $end\n")
-        for module in sorted(vcd['sigs'].keys()):
-            fh.write("  $scope module %s $end\n" % module)
-            for sig in sorted(vcd['sigs'][module].keys()):
-                code = vcd['sigs'][module][sig]
-                fh.write("   $var wire 32 v%x %s [31:0] $end\n" % (code, sig))
-                all_codes[code] = 1
-            fh.write("  $upscope $end\n")
+        writeScope(scopeSigs)
        fh.write(" $upscope $end\n")
        fh.write("$enddefinitions $end\n")
        fh.write("\n")

        first = True
-        for time in sorted(vcd['values']):
+        for time in sorted(values):
            if first:
                first = False
                # Start with Z for any signals without time zero data
-                for code in sorted(all_codes.keys()):
-                    if code not in vcd['values'][time]:
-                        vcd['values'][time][code] = None
+                for code in sorted(all_codes):
+                    if code not in values[time]:
+                        addValue(code, time, None)
            fh.write("#%d\n" % time)
-            for code in sorted(vcd['values'][time].keys()):
-                value = vcd['values'][time][code]
+            for code in sorted(values[time]):
+                value = values[time][code]
                if value is None:
                    fh.write("bz v%x\n" % code)
+                elif isinstance(value, str):
+                    fh.write("b%s v%x\n" % (value, code))
                else:
                    fh.write("b%s v%x\n" % (format(value, 'b'), code))

@@ -484,7 +546,7 @@ Verilator_gantt creates a visual representation to help analyze Verilator
 For documentation see
 https://verilator.org/guide/latest/exe_verilator_gantt.html""",
    epilog=
-    """Copyright 2018-2023 by Wilson Snyder. This program is free software; you
+    """Copyright 2018-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -504,7 +566,8 @@ parser.add_argument('filename',

 Args = parser.parse_args()

-process(Args.filename)
+read_data(Args.filename)
+report()
 if not Args.no_vcd:
    write_vcd(Args.vcd)

@@ -1,7 +1,7 @@
 #!/usr/bin/env python3
 # pylint: disable=C0114,C0209
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify the Verilator internals under the terms
 # of either the GNU Lesser General Public License Version 3 or the Perl
 # Artistic License Version 2.0.
@@ -181,7 +181,7 @@ in each Verilog block.
 For documentation see
 https://verilator.org/guide/latest/exe_verilator_profcfunc.html""",
    epilog=
-    """Copyright 2002-2023 by Wilson Snyder. This program is free software; you
+    """Copyright 2002-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -26,4 +26,5 @@ if git log --format=%B -n 1 "$COMMIT" | grep -q -i '\[CI\s\+ccache\s\+clear\]';
 fi

 # Dump stats, then zero stats
+# 22.04: ccache -s -v -z
 ccache -s -z
@@ -61,7 +61,7 @@ if [ "$CI_BUILD_STAGE_NAME" = "build" ]; then
    sudo apt-get update
    sudo apt-get install ccache help2man libfl-dev ||
    sudo apt-get install ccache help2man libfl-dev
-    if [ "$CI_RUNS_ON" != "ubuntu-22.04" ]; then
+    if [ "$CI_RUNS_ON" = "ubuntu-20.04" ]; then
      # Some conflict of libunwind verison on 22.04, can live without it for now
      sudo apt-get install libgoogle-perftools-dev ||
      sudo apt-get install libgoogle-perftools-dev
@@ -70,6 +70,10 @@ if [ "$CI_BUILD_STAGE_NAME" = "build" ]; then
      sudo apt-get install libsystemc libsystemc-dev ||
      sudo apt-get install libsystemc libsystemc-dev
    fi
+    if [ "$CI_RUNS_ON" = "ubuntu-22.04" ]; then
+      sudo apt-get install bear mold ||
+      sudo apt-get install bear mold
+    fi
    if [ "$COVERAGE" = 1 ]; then
      yes yes | sudo cpan -fi Parallel::Forker
    fi
@@ -102,8 +106,8 @@ elif [ "$CI_BUILD_STAGE_NAME" = "test" ]; then
    sudo apt-get install gdb gtkwave lcov libfl-dev ccache
    # Required for test_regress/t/t_dist_attributes.pl
    if [ "$CI_RUNS_ON" = "ubuntu-22.04" ]; then
-      sudo apt-get install libclang-dev ||
-      sudo apt-get install libclang-dev
+      sudo apt-get install libclang-dev mold ||
+      sudo apt-get install libclang-dev mold
      pip3 install clang==14.0
    fi
    if [ "$CI_RUNS_ON" = "ubuntu-20.04" ] || [ "$CI_RUNS_ON" = "ubuntu-22.04" ]; then
@@ -49,6 +49,7 @@ if [ "$CI_BUILD_STAGE_NAME" = "build" ]; then
    ./configure --enable-longtests --enable-ccwarn ${CI_M32:+--enable-m32}
    ccache -z
    "$MAKE" -j "$NPROC" -k
+    # 22.04: ccache -s -v
    ccache -s
    if [ "$CI_OS_NAME" = "osx" ]; then
      file bin/verilator_bin
@@ -89,8 +90,8 @@ elif [ "$CI_BUILD_STAGE_NAME" = "test" ]; then
    export VERILATOR_TEST_NO_GPROF=1 # gprof is a bit different on FreeBSD, disable
  fi

-  # Run sanitize on Ubuntu 20.04 only
-  [ "$CI_RUNS_ON" = 'ubuntu-20.04' ] && [ "$CI_M32" = "" ] && sanitize='--sanitize' || sanitize=''
+  # Run sanitize on Ubuntu 22.04 only
+  [ "$CI_RUNS_ON" = 'ubuntu-22.04' ] && [ "$CI_M32" = "" ] && sanitize='--sanitize' || sanitize=''

  # Run the specified test
  ccache -z
@@ -180,6 +181,7 @@ elif [ "$CI_BUILD_STAGE_NAME" = "test" ]; then
      fatal "Unknown test: $TESTS"
      ;;
  esac
+  # 22.04: ccache -s -v
  ccache -s

  # Upload coverage data
@@ -4,12 +4,12 @@ if (-Not (Test-Path $PWD/../.ccache/win_bison.exe)) {
 	mkdir build
 	cd build
 	cmake .. --install-prefix $PWD/../../../.ccache
-	cmake --build . --config Release
+	cmake --build . --config Release -j 3
 	cmake --install . --prefix $PWD/../../../.ccache
 	cd ../..
 }
 mkdir build
 cd build
 cmake .. --install-prefix $PWD/../install
-cmake --build . --config Release
+cmake --build . --config Release -j 3
 cmake --install . --prefix $PWD/../install
@@ -4,7 +4,7 @@ cd examples/cmake_tracing_c
 mkdir build
 cd build
 cmake ..
-cmake --build . --config Release
+cmake --build . --config Release -j 3
 Release/example.exe
 cd ..
 Remove-Item -path build -recurse
@@ -8,6 +8,16 @@

 FROM ubuntu:22.04

+# Create the user
+RUN groupadd verilator \
+    && useradd -g verilator -m verilator -s /bin/bash \
+    && apt-get update \
+    && apt-get install --no-install-recommends -y sudo \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/* \
+    && echo verilator ALL=\(root\) NOPASSWD:ALL > /etc/sudoers.d/verilator \
+    && chmod 0440 /etc/sudoers.d/verilator
+
 RUN apt-get update \
    && DEBIAN_FRONTEND=noninteractive \
       apt-get install --no-install-recommends -y \
@@ -33,7 +43,6 @@ RUN apt-get update \
                        perl \
                        python3 \
                        wget \
-                        zlibc \
                        zlib1g \
                        zlib1g-dev \
    && apt-get clean \
@@ -52,6 +61,8 @@ COPY build.sh /tmp/build.sh

 ENV VERILATOR_AUTHOR_SITE=1

+USER verilator
+
 WORKDIR /work

 ENTRYPOINT [ "/tmp/build.sh" ]
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 Verilator Build Docker Container
@@ -19,10 +19,16 @@ RUN apt-get update \
                        ccache \
                        flex \
                        git \
+                        help2man \
+                        libfl2 \
                        libfl-dev \
                        libgoogle-perftools-dev \
+                        numactl \
                        perl \
+                        perl-doc \
                        python3 \
+                        zlib1g \
+                        zlib1g-dev \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

@@ -39,7 +45,7 @@ RUN git clone "${REPO}" verilator && \
    git checkout "${SOURCE_COMMIT}" && \
    autoconf && \
    ./configure && \
-    make -j "$(nproc)" && \
+    make && \
    make install && \
    cd .. && \
    rm -r verilator
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 Verilator Executable Docker Container
@@ -1,6 +1,6 @@
 # DESCRIPTION: codecov.io config
 #
-# Copyright 2020-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2020-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
@@ -1,20 +1,20 @@
 # DESCRIPTION: Process this file with autoconf to produce a configure script.
 #
-# Copyright 2003-2023 by Wilson Snyder. Verilator is free software; you
+# Copyright 2003-2024 by Wilson Snyder. Verilator is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU Lesser
 # General Public License Version 3 or the Perl Artistic License Version 2.0.
 # SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

-# When releasing, also update header of Changes file
+# When releasing, also update header of Changes file, and CmakeLists.txt,
 # and commit using "devel release" or "Version bump" message
 # Then 'make maintainer-dist'
 #AC_INIT([Verilator],[#.### YYYY-MM-DD])
 #AC_INIT([Verilator],[#.### devel])
-AC_INIT([Verilator],[5.006 2023-01-22],
+AC_INIT([Verilator],[5.020 2024-01-01],
        [https://verilator.org],
        [verilator],[https://verilator.org])

-AC_CONFIG_HEADERS(src/config_build.h)
+AC_CONFIG_HEADERS(src/config_package.h)
 AC_CONFIG_FILES(Makefile src/Makefile src/Makefile_obj include/verilated.mk include/verilated_config.h verilator.pc verilator-config.cmake verilator-config-version.cmake)

 # Version
@@ -24,6 +24,10 @@ AC_SUBST(PACKAGE_VERSION_NUMBER)
 AC_DEFINE_UNQUOTED([PACKAGE_VERSION_NUMBER_STRING],["$PACKAGE_VERSION_NUMBER"],[Package version as a number])
 VERILATOR_VERSION_INTEGER=`AS_ECHO("$PACKAGE_VERSION") | [sed 's/\([0-9]\)\.\([0-9][0-9][0-9]\) .*/\1\2000/g']`
 AC_SUBST(VERILATOR_VERSION_INTEGER)
+AC_DEFINE_UNQUOTED([PACKAGE_VERSION_STRING_CHAR],
+        [static const char* const PACKAGE_STRING_UNUSED = "$PACKAGE_STRING";],
+        [Package version as a number])
+AC_SUBST(PACKAGE_VERSION_STRING_CHAR)

 # Ignore automake flags passed by Ubuntu builds
 AC_ARG_ENABLE([dependency-tracking],
@@ -68,7 +72,7 @@ AC_MSG_CHECKING(whether to use -m32)
 AC_ARG_ENABLE([m32],
              [AS_HELP_STRING([--enable-m32],
                              [Use -m32 for all compilation and link,
-                               including Verialtor and generated models.])],
+                               including Verilator and generated models.])],
              [case "${enableval}" in
                yes) CFG_ENABLE_M32=yes ;;
                no)  CFG_ENABLE_M32=no ;;
@@ -350,7 +354,7 @@ AC_SUBST(CFG_CXXFLAGS_PROFILE)
 # Macros work such that first option that passes is the one we take
 # Currently enable c++17/c++14 due to packaged SystemC dependency
 # c++17 is the newest that Verilator is regularly tested to support
-# c++11 is the oldest that Verilator supports
+# c++14 is the oldest that Verilator supports
 # gnu is required for Cygwin to compile verilated.h successfully
 #_MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=gnu++20)
 #_MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=c++20)
@@ -358,8 +362,6 @@ _MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=gnu++17)
 _MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=c++17)
 _MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=gnu++14)
 _MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=c++14)
-_MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=gnu++11)
-_MY_CXX_CHECK_SET(CFG_CXXFLAGS_STD_NEWEST,-std=c++11)
 AC_SUBST(CFG_CXXFLAGS_STD_NEWEST)

 # Flags for compiling Verilator internals including parser, and Verilated files
@@ -371,10 +373,9 @@ _MY_CXX_CHECK_OPT(CFG_CXXFLAGS_WEXTRA,-Wthread-safety)
 AC_SUBST(CFG_CXXFLAGS_WEXTRA)

 # Flags for coroutine support for dynamic scheduling
-_MY_CXX_CHECK_IFELSE(
-  -fcoroutines-ts,
-  [CFG_CXXFLAGS_COROUTINES="-fcoroutines-ts"],
-  [CFG_CXXFLAGS_COROUTINES="-fcoroutines"])
+_MY_CXX_CHECK_SET(CFG_CXXFLAGS_COROUTINES,-fcoroutines-ts)
+_MY_CXX_CHECK_SET(CFG_CXXFLAGS_COROUTINES,-fcoroutines)
+_MY_CXX_CHECK_SET(CFG_CXXFLAGS_COROUTINES,-fcoroutines-ts -Wno-deprecated-experimental-coroutine)
 AC_SUBST(CFG_CXXFLAGS_COROUTINES)

 # HAVE_COROUTINES
@@ -442,14 +443,19 @@ m4_foreach([cflag],[
        [-mno-cet],
        [-Qunused-arguments],
        [-Wno-bool-operation],
-        [-Wno-tautological-bitwise-compare],
+        [-Wno-c++11-narrowing],
+        [-Wno-constant-logical-operand],
+        [-Wno-non-pod-varargs],
+        [-Wno-overloaded-virtual],
        [-Wno-parentheses-equality],
+        [-Wno-shadow],
        [-Wno-sign-compare],
+        [-Wno-tautological-bitwise-compare],
        [-Wno-uninitialized],
+        [-Wno-unused-but-set-parameter],
        [-Wno-unused-but-set-variable],
        [-Wno-unused-parameter],
-        [-Wno-unused-variable],
-        [-Wno-shadow]],[
+        [-Wno-unused-variable]],[
    _MY_CXX_CHECK_OPT(CFG_CXXFLAGS_NO_UNUSED,cflag)
    # CMake will test what flags work itself, so pass all flags through to it
    CFG_CXX_FLAGS_CMAKE="$CFG_CXX_FLAGS_CMAKE cflag"
@@ -470,6 +476,10 @@ m4_foreach([ldflag], [
 AC_SUBST(CFG_LDLIBS_THREADS)
 AC_SUBST(CFG_LDFLAGS_THREADS_CMAKE)

+# If 'mold' is installed, use it to link for faster buildtimes
+_MY_LDLIBS_CHECK_OPT(CFG_LDFLAGS_SRC, -fuse-ld=mold)
+_MY_LDLIBS_CHECK_OPT(CFG_LDFLAGS_VERILATED, -fuse-ld=mold)
+
 # When linking partially statically
 if test "$CFG_ENABLE_PARTIAL_STATIC" = "yes"; then
  _MY_LDLIBS_CHECK_OPT(CFG_LDFLAGS_SRC, -static-libgcc)
@@ -480,6 +490,7 @@ else
  LTCMALLOC=-ltcmalloc_minimal
 fi
 AC_SUBST(CFG_LDFLAGS_SRC)
+AC_SUBST(CFG_LDFLAGS_VERILATED)

 # The pthread library is required by tcmalloc, so add it if it exists. If it
 # does not, the tcmalloc check below will fail anyway, and linking against
@@ -508,14 +519,14 @@ _MY_LDLIBS_CHECK_IFELSE(
   fi])
 AC_SUBST(CFG_LIBS)

-# Need C++11 at least
-#Alternative: AX_CXX_COMPILE_STDCXX([11])
+# Need C++14 at least
+#Alternative: AX_CXX_COMPILE_STDCXX([14])
 AC_DEFUN([_MY_CXX_CHECK_CXX_VER],
-   [# _MY_CXX_CHECK_CXX_VER(flag) -- Check if compiler runs C++11
+   [# _MY_CXX_CHECK_CXX_VER(flag) -- Check if compiler runs C++14
    # Set $_my_result
    AC_LINK_IFELSE(
     [AC_LANG_PROGRAM([#include <thread>
-#if (__cplusplus < 201103L)
+#if (__cplusplus < 201402L)
 # error "Too old"
 #endif
    ], [[ ]])],
@@ -531,13 +542,13 @@ AC_DEFUN([_MY_CXX_CHECK_CXX_VER],
 # Add $CFG_CXXFLAGS_STD only if can't compile correctly otherwise,
 # as adding std= when not needed can cause errors with the C++ std library.
 CFG_CXXFLAGS_STD=$CFG_CXXFLAGS_STD_NEWEST
-AC_MSG_CHECKING(whether $CXX supports C++11)
+AC_MSG_CHECKING(whether $CXX supports C++14)
 _MY_CXX_CHECK_CXX_VER()
 AC_MSG_RESULT($_my_result)
 if test "$_my_result" = "no" ; then
   CXXFLAGS="$CXXFLAGS $CFG_CXXFLAGS_STD"
   CFG_CXX_FLAGS_CMAKE="$CFG_CXX_FLAGS_CMAKE $CFG_CXXFLAGS_STD"
-   AC_MSG_CHECKING(whether $CXX supports C++11 with $CFG_CXXFLAGS_STD)
+   AC_MSG_CHECKING(whether $CXX supports C++14 with $CFG_CXXFLAGS_STD)
   _MY_CXX_CHECK_CXX_VER()
   AC_MSG_RESULT($_my_result)
 else
@@ -547,13 +558,28 @@ else
 fi
 if test "$_my_result" = "no" ; then
   AC_MSG_NOTICE([[]])
-   AC_MSG_ERROR([[the $CXX compiler appears to not support C++11.
+   AC_MSG_ERROR([[the $CXX compiler appears to not support C++14.

-Verilator requires a C++11 or newer compiler.]])
+Verilator requires a C++14 or newer compiler.]])

 fi
 AC_SUBST(CFG_CXXFLAGS_STD)

+# Compiler precompiled header options (assumes either gcc or clang++)
+AC_MSG_CHECKING([for $CXX precompile header include option])
+if $CXX --help | grep include-pch >/dev/null 2>/dev/null ; then
+   # clang
+   CFG_CXXFLAGS_PCH_I=-include-pch
+   CFG_GCH_IF_CLANG=.gch
+else
+   # GCC
+   CFG_CXXFLAGS_PCH_I=-include
+   CFG_GCH_IF_CLANG=
+fi
+AC_MSG_RESULT($CFG_CXXFLAGS_PCH_I)
+AC_SUBST(CFG_CXXFLAGS_PCH_I)
+AC_SUBST(CFG_GCH_IF_CLANG)
+
 # Checks for library functions.
 AC_CHECK_MEMBER([struct stat.st_mtim.tv_nsec],
    [AC_DEFINE([HAVE_STAT_NSEC],[1],[Defined if struct stat has st_mtim.tv_nsec])],
@@ -568,7 +594,7 @@ ACO_SAVE_LIBS="$LIBS"
 LIBS="$LIBS -lsystemc"
 AC_LINK_IFELSE(
    [AC_LANG_PROGRAM([[#include <systemc.h>
-                     extern "C" int sc_main(int argc, char* argv[]) {}
+                     extern "C" int sc_main(int argc, char* argv[]) { return 0; }
                     ]],[[sc_version()]])],
    [_my_result=yes
     AC_DEFINE([HAVE_SYSTEMC],[1],[Defined if have SystemC library])],
@@ -10,9 +10,12 @@ contributions flow more efficiently.
 Did you find a bug?
 -------------------

- Please **Ensure the bug was not already reported** by searching
+- Please **ensure the bug was not already reported** by searching
  `Verilator Issues <https://verilator.org/issues>`__.

+- Please **download the latest development GitHub version**, build, and see
+  if the issue has been fixed.
+
 - If you're unable to find an open issue addressing the problem, `open a
  new Verilator issue <https://verilator.org/issues/new>`__.

@@ -20,21 +23,28 @@ Did you find a bug?
    demonstrating the bug and expected behavior that is not occurring.

  - The ideal example works against other simulators, and is in the
-    test_regress/t test format, as described in `docs/internals
-    <internals.rst>`__.
+    test_regress/t test format, as described in
+    `Verilator Internals Documentation
+    <https://github.com/verilator/verilator/blob/master/docs/internals.rst>`__.


 Did you write a patch that fixes a bug?
 ---------------------------------------

- Please `Open a new issue <https://verilator.org/issues/new>`__.
+- Please `Open a new Verilator issue <https://verilator.org/issues/new>`__
+  if there is not one already describing the bug.

- You may attach a patch to the issue, or (preferred) may request a
-  GitHub pull request.
+- Please `Open a Verilator pull request
+  <https://github.com/verilator/verilator/pulls>`__.

-  - Verilator uses GitHub Actions to provide continuous integration.  You
-    may want to enable Actions on your GitHub branch to ensure your changes
-    keep the tests passing. See `docs/internals <internals.rst>`__.
+- See the coding conventions, and other developer information in
+  ``docs/internals.rst`` in the distribution, or as rendered at
+  `Verilator Internals Documentation
+  <https://github.com/verilator/verilator/blob/master/docs/internals.rst>`__.
+
+- Verilator uses GitHub Actions to provide continuous integration.  You
+  may want to enable Actions on your GitHub branch to ensure your changes
+  keep the tests passing.

 - Your source-code contributions must be certified as open source,
  under the `Developer Certificate of
@@ -56,13 +66,10 @@ Did you write a patch that fixes a bug?
    contribution.

 - Your test contributions are generally considered released into the
-  Creative Commons Public Domain (CC0), unless you request otherwise or
+  Creative Commons Public Domain (CC0), unless you request otherwise, or
  put a GNU/Artistic license on your file.

- Most important is we get your patch. If you'd like to clean up
-  indentation and related issues ahead of our feedback, that is
-  appreciated; please see the coding conventions in `docs/internals
-  <internals.rst>`__.
+- Most important is we get your patch.


 Do you have questions?
@@ -3,6 +3,7 @@ under the Developer Certificate of Origin (https://developercertificate.org/).

 Please see the Verilator manual for 200+ additional contributors. Thanks to all.

+Adam Bagley
 Adrien Le Masle
 Ahmed El-Mahmoudy
 Aleksander Kiryk
@@ -11,24 +12,38 @@ Alex Chadwick
 Aliaksei Chapyzhenka
 Ameya Vikram Singh
 Andreas Kuster
+Andrei Kostovski
+Andrew Miloradovsky
 Andrew Nolte
+Anthony Donlon
 Arkadiusz Kozdra
+Aylon Chaim Porat
+Cameron Kirk
+Chih-Mao Chen
 Chris Randall
 Chuxuan Wang
 Conor McCullough
 Dan Petrisko
 Daniel Bates
 David Horton
+David Ledger
 David Metz
 David Stanford
 David Turner
+Don Williamson
 Drew Ranck
 Drew Taussig
 Driss Hafdi
 Edgar E. Iglesias
 Eric Rippey
+Ethan Sifferman
+Eyck Jentzsch
 Fan Shupei
+february cozzocrea
+Felix Neumärker
 Felix Yan
+Frans Skarman
+G-A. Kamendje
 Garrett Smith
 Geza Lore
 Gianfranco Costamagna
@@ -36,17 +51,19 @@ Glen Gibb
 Gökçe Aydos
 Graham Rushton
 Guokai Chen
+Gus Smith
 Gustav Svensk
-G-A. Kamendje
 Harald Heckmann
+Hennadii Chernyshchyk
 Howard Su
 Huang Rui
 Huanghuang Zhou
 HungMingWu
 HyungKi Jeong
+Ícaro Lima
+Ilya Barkov
 Iru Cai
 Ivan Vnučec
-Ilya Barkov
 Iztok Jeras
 Jake Merdich
 James Hanlon
@@ -58,16 +75,26 @@ Jamie Iles
 Jan Van Winkel
 Jean Berniolles
 Jeremy Bennett
+Jesse Taube
 Jevin Sweval
 Jiacheng Qian
+Jiamin Zhu
+Jinyan Xu
 Jiuyang Liu
 Joey Liu
 John Coiner
 John Demme
+John Wehle
 Jonathan Drolet
+Jordan McConnon
+Jose Loyola
+Josep Sans
+Joseph Nwabueze
 Josh Redford
+Julian Daube
 Julie Schwartz
 Julien Margetts
+Justin Thiel
 Kaleb Barrett
 Kamil Rakoczy
 Kanad Kanhere
@@ -75,6 +102,8 @@ Keith Colbert
 Kevin Kiningham
 Kritik Bhimani
 Krzysztof Bieganski
+Krzysztof Boronski
+Krzysztof Boroński
 Kuba Ober
 Larry Doolittle
 Ludwig Rogiers
@@ -92,6 +121,7 @@ Martin Stadler
 Matthew Ballance
 Michael Killough
 Michaël Lefebvre
+Michal Czyz
 Mike Popoloski
 Miodrag Milanović
 Mladen Slijepcevic
@@ -100,9 +130,11 @@ Mostafa Gamal
 Nandu Raj
 Nathan Kohagen
 Nathan Myers
+Oleh Maksymenko
 Patrick Stewart
 Paul Wright
 Pawel Sagan
+Peter Debacker
 Peter Horvath
 Peter Monsson
 Philipp Wagner
@@ -110,26 +142,34 @@ Pierre-Henri Horrein
 Pieter Kapsenberg
 Piotr Binkowski
 Qingyao Sun
+Quentin Corradi
 Rafal Kapuscik
+Rasfunk
 Raynard Qiao
 Richard Myers
+Risto Pejašinović
+Robert Balas
 Rupert Swarbrick
 Ryszard Rozak
 Samuel Riedel
 Sean Cross
 Sebastien Van Cauwenberghe
 Sergi Granell
+Srinivasan Venkataramanan
 Stefan Wallentowitz
 Stephen Henry
 Steven Hugg
 Sören Tempel
 Teng Huang
+Tim Hutt
 Tim Snyder
 Tobias Rosenkranz
 Tobias Wölfel
 Todd Strader
 Tomasz Gorochowik
 Topa Topino
+Toru Niina
+Tudor Timi
 Tymoteusz Blazejczyk
 Udi Finkelstein
 Unai Martinez-Corral
@@ -137,14 +177,18 @@ Varun Koyyalagunta
 Vassilis Papaefstathiou
 Veripool API Bot
 Victor Besyakov
+Vito Gamberini
 William D. Jones
 Wilson Snyder
 Xi Zhang
 Yinan Xu
 Yoda Lee
 Yossi Nivin
+Yu-Sheng Lin
 Yuri Victorovich
 Yutetsu TAKATSUKASA
-Yu-Sheng Lin
 Yves Mathieu
 Zhanglei Wang
+Zhou Shen
+Zixi Li
+أحمد المحمودي
@@ -5,7 +5,7 @@
 #
 # Code available from: https://verilator.org
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
@@ -18,12 +18,16 @@
 RST2HTML = rst2html
 PYTHON3 = python3
 DOXYGEN = doxygen
-SPHINXOPTS ?= -c guide
 SPHINXBUILD ?= sphinx-build

 SOURCEDIR = guide
 BUILDDIR = _build

+SPHINXOPTS ?= -c guide
+ifneq ($(VERILATOR_ANALYTICS_ID),)
+SPHINXOPTS += -D html_theme_options.analytics_id=$(VERILATOR_ANALYTICS_ID)
+endif
+
 ######################################################################

 .SUFFIXES:
@@ -0,0 +1,21 @@
+Verilator Documentation
+=======================
+
+This folder contains sources for Verilator documentation.
+
+For formatted documentation see:
+
+- `Verilator README <https://github.com/verilator/verilator>`_
+
+- `Verilator installation and package directory structure
+  <https://verilator.org/install>`_
+
+- `Verilator manual (HTML) <https://verilator.org/verilator_doc.html>`_,
+  or `Verilator manual (PDF) <https://verilator.org/verilator_doc.pdf>`_
+
+- `Subscribe to Verilator announcements
+  <https://github.com/verilator/verilator-announce>`_
+
+- `Verilator forum <https://verilator.org/forum>`_
+
+- `Verilator issues <https://verilator.org/issues>`_
@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-# pylint: disable=C0112,C0114,C0115,C0116,C0209,C0301,R0201,R0903
+# pylint: disable=C0112,C0114,C0115,C0116,C0209,C0301,R0903
 # -*- Python -*- See copyright, etc below
 ######################################################################

@@ -40,7 +40,7 @@ parser = argparse.ArgumentParser(
    formatter_class=argparse.RawDescriptionHelpFormatter,
    description="""Read a file and extract documentation data.""",
    epilog=
-    """ Copyright 2021-2023 by Wilson Snyder.  This package is free software;
+    """ Copyright 2021-2024 by Wilson Snyder.  This package is free software;
 you can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -55,7 +55,7 @@ parser = argparse.ArgumentParser(
    formatter_class=argparse.RawDescriptionHelpFormatter,
    description="""Post-process Sphinx HTML.""",
    epilog=
-    """ Copyright 2021-2023 by Wilson Snyder.  This package is free software;
+    """ Copyright 2021-2024 by Wilson Snyder.  This package is free software;
 you can redistribute it and/or modify it under the terms of either the GNU
 Lesser General Public License Version 3 or the Perl Artistic License
 Version 2.0.
@@ -1,4 +1,4 @@
-.. comment: generated by t_lint_width_docs_bad
+.. comment: generated by t_lint_widthexpand_docs_bad
 .. code-block:: sv
   :linenos:
   :emphasize-lines: 3
@@ -1,4 +1,4 @@
-.. comment: generated by t_lint_width_docs_bad
+.. comment: generated by t_lint_widthexpand_docs_bad
 .. code-block:: sv
   :emphasize-lines: 1

@@ -0,0 +1,4 @@
+.. comment: generated by t_lint_widthexpand_docs_bad
+.. code-block::
+
+   %Warning-WIDTHEXPAND: example.v:3:29 Bit extraction of array[4:0] requires 3 bit index, not 2 bits.
@@ -1,4 +0,0 @@
-.. comment: generated by t_lint_width_docs_bad
-.. code-block::
-
-   %Warning-WIDTH: example.v:3:29 Bit extraction of array[4:0] requires 3 bit index, not 2 bits.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 ****************
@@ -1,7 +1,7 @@
 # pylint: disable=C0103,C0114,C0116,C0301,E0402,W0622
 #
 # Configuration file for Verilator's Sphinx documentation builder.
-# Copyright 2003-2023 by Wilson Snyder.
+# Copyright 2003-2024 by Wilson Snyder.
 # SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0
 #
 # This file only contains overridden options. For a full list:
@@ -10,7 +10,6 @@
 # ----------------------------------------------------------------------
 # -- Path setup

-from datetime import datetime
 import os
 import re
 import sys
@@ -24,10 +23,18 @@ def get_vlt_version():
    filename = "../../Makefile"
    with open(filename, "r", encoding="utf8") as fh:
        for line in fh:
-            match = re.search(r"PACKAGE_VERSION_NUMBER *= *([a-z0-9.]+)", line)
+            match = re.search(r"PACKAGE_VERSION *= *([a-z0-9.]+) +([-0-9]+)",
+                              line)
            if match:
-                return match.group(1)
-    return "unknown"
+                return match.group(1), match.group(2)
+            match = re.search(r"PACKAGE_VERSION *= *([a-z0-9.]+) +devel", line)
+            if match:
+                try:
+                    data = os.popen('git log -n 1 --pretty=%cs').read()
+                except Exception:  # pylint: disable=broad-except
+                    data = ""  # fallback, and Sphinx will fill in today's date
+                return "Devel " + match.group(1), data
+    return "unknown", "unknown"


 def setup(app):
@@ -44,8 +51,8 @@ author = 'Wilson Snyder'
 # The master toctree document.
 master_doc = "index"

-version = get_vlt_version()
-release = get_vlt_version()
+version, today_fmt = get_vlt_version()
+release = version

 rst_prolog = """
 .. role:: vlopt(option)
@@ -89,9 +96,6 @@ source_suffix = [".rst"]
 # Add any paths that contain templates here, relative to this directory.
 templates_path = ['_templates']

-# Date format to ISO
-today_fmt = datetime.now().strftime("%F")
-
 # If true, `todo` and `todoList` produce output, else they produce nothing.
 todo_include_todos = True

@@ -111,7 +115,6 @@ html_theme = 'sphinx_rtd_theme'
 html_theme_path = [sphinx_rtd_theme.get_html_theme_path()]

 html_theme_options = {
-    'analytics_id': 'G-D27B0C9QEB',
    'logo_only': True,
    'style_nav_header_background': "#45acf8",  # Default is #2980B9
    # 'canonical_url':
@@ -227,4 +230,4 @@ spelling_ignore_contributor_names = True

 breathe_projects = {"verilated": "_build/doxygen/verilated/xml/"}
 breathe_default_project = "verilated"
-breathe_default_members = ('members')
+breathe_default_members = 'members'
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Connecting:
@@ -120,6 +120,52 @@ code.  This is a feature, as using the SystemC pin interconnect scheme
 everywhere would reduce performance by an order of magnitude.


+Verilated API
+=============
+
+The API to a Verilated model is the C++ headers in the include/ directory
+in the distribution.  These headers use Doxygen comments, `///` and `//<`,
+to indicate and document those functions that are part of the Verilated
+public API.
+
+Process-Level Clone APIs
+--------------------------
+
+Modern operating systems support process-level clone (a.k.a copying, forking)
+with system call interfaces in C/C++, e.g., :code:`fork()` in Linux.
+
+However, after cloning a parent process, some resources cannot be inherited
+in the child process. For example, in POSIX systems, when you fork a process,
+the child process inherits all the memory of the parent process. However,
+only the thread that called fork is replicated in the child process. Other
+threads are not.
+
+Therefore, to support the process-level clone mechanisms, Verilator supports
+:code:`prepareClone()` and :code:`atClone()` APIs to allow the user to manually
+re-construct the model in the child process. The two APIs handle all necessary
+resources required for releasing and re-initializing before and after cloning.
+
+The two APIs are supported in the verilated models. Here is an example of usage
+with Linux :code:`fork()` and :code:`pthread_atfork` APIs:
+
+.. code-block:: C++
+
+    // static function pointers to fit pthread_atfork
+    static auto prepareClone = [](){ topp->prepareClone(); };
+    static auto atClone = [](){ topp->atClone(); };
+
+    // in main function, register the handlers:
+    pthread_atfork(prepareClone, atClone, atClone);
+
+For better flexibility, you can also manually call the handlers before and
+after :code:`fork()`.
+
+With the process-level clone APIs, users can create process-level snapshots
+for the verilated models. While the Verilator save/restore option provides
+persistent and circuit-dependent snapshots, the process-level clone APIs
+enable in-memory, circuit-transparent, and highly efficient snapshots.
+
+
 Direct Programming Interface (DPI)
 ==================================

@@ -374,9 +420,11 @@ changed on the specified clock edge.
 .. code-block:: bash

     cat >our.v <<'EOF'
-       module our (input clk);
-          reg readme   /*verilator public_flat_rd*/;
-          reg writeme  /*verilator public_flat_rw @(posedge clk) */;
+       module our #(
+          parameter WIDTH /*verilator public_flat_rd*/ = 32
+       ) (input clk);
+          reg [WIDTH-1:0] readme   /*verilator public_flat_rd*/;
+          reg [WIDTH-1:0] writeme  /*verilator public_flat_rw @(posedge clk) */;
          initial $finish;
       endmodule
     EOF
@@ -398,12 +446,14 @@ accesses the above signal "readme" would be:
           vpiHandle vh1 = vpi_handle_by_name((PLI_BYTE8*)"TOP.our.readme", NULL);
           if (!vh1) vl_fatal(__FILE__, __LINE__, "sim_main", "No handle found");
           const char* name = vpi_get_str(vpiName, vh1);
-           printf("Module name: %s\n", name);  // Prints "readme"
+           const char* type = vpi_get_str(vpiType, vh1);
+           const int size = vpi_get(vpiSize, vh1);
+           printf("register name: %s, type: %s, size: %d\n", name, type, size);  // Prints "register name: readme, type: vpiReg, size: 32"

           s_vpi_value v;
           v.format = vpiIntVal;
           vpi_get_value(vh1, &v);
-           printf("Value of v: %d\n", v.value.integer);  // Prints "readme"
+           printf("Value of %s: %d\n", name, v.value.integer);  // Prints "Value of readme: 0"
       }

       int main(int argc, char** argv) {
@@ -479,11 +529,12 @@ distribution.
 Verilated and VerilatedContext
 ==============================

-Multiple Verilated models may be part of the same simulation context, that
-is share a VPI interface, sense of time, and common settings.  This common
-simulation context information is stored in a ``VerilatedContext``
+Multiple C++ Verilated models may be part of the same simulation context,
+that is share a VPI interface, sense of time, and common settings.  This
+common simulation context information is stored in a ``VerilatedContext``
 structure.  If a ``VerilatedContext`` is not created prior to creating a
-model, a default global one is created automatically.
+model, a default global one is created automatically.  SystemC requires
+using only the single, default VerilatedContext.

 The ``Verilated::`` methods, including the ``Verilated::commandArgs`` call
 shown above, call VerilatedContext methods using the default global
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 *******************************
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 ************************
@@ -27,107 +27,137 @@ Alliance <https://chipsalliance.org>`_, `Antmicro Ltd
 <https://antmicro.com>`_ and `Shunyao CAD <https://shunyaocad.com>`_.

 Previous major corporate sponsors of Verilator, by providing significant
-contributions of time or funds include: Atmel Corporation, Cavium
-Inc., Compaq Corporation, Digital Equipment Corporation, Embecosm Ltd.,
-Hicamp Systems, Intel Corporation, Mindspeed Technologies Inc., MicroTune
-Inc., picoChip Designs Ltd., Sun Microsystems Inc., Nauticus Networks Inc.,
-SiCortex Inc, and Shunyao CAD.
+contributions of time or funds include: Antmicro Ltd., Atmel Corporation,
+Compaq Corporation, Digital Equipment Corporation, Embecosm Ltd., Hicamp
+Systems, Intel Corporation, Marvell Inc., Mindspeed Technologies Inc.,
+MicroTune Inc., picoChip Designs Ltd., Sun Microsystems Inc., Nauticus
+Networks Inc., SiCortex Inc, Shunyao CAD, and Western Digital Inc.

-The contributors of major functionality are: Krzysztof
-Bieganski, Byron Bradley, Jeremy Bennett, Lane Brooks, John Coiner, Duane
-Galbi, Geza Lore, Todd Strader, Stefan Wallentowitz, Paul Wasson, Jie Xu,
-and Wilson Snyder.  Major testers included Jeff Dutton, Jonathon Donaldson,
-Ralf Karge, David Hewson, Iztok Jeras, Wim Michiels, Alex Solomatnikov,
-Sebastien Van Cauwenberghe, Gene Weber, and Clifford Wolf.
+The contributors of major functionality are: Jeremy Bennett, Krzysztof
+Bieganski, Byron Bradley, Lane Brooks, John Coiner, Duane Galbi, Geza Lore,
+Todd Strader, Yutetsu Takatsukasa, Stefan Wallentowitz, Paul Wasson, Jie
+Xu, and Wilson Snyder.

 Some of the people who have provided ideas, and feedback for Verilator
 include:

-David Addison, Tariq B. Ahmad, Nikana Anastasiadis, Vasu Arasanipalai, Jens
-Arm, Sharad Bagri, Matthew Ballance, Andrew Bardsley, Matthew Barr, Geoff
-Barrett, Kaleb Barrett, Julius Baxter, Jeremy Bennett, Michael Berman, Jean
-Berniolles, Victor Besyakov, Moinak Bhattacharyya, Krzysztof Bieganski,
-David Binderman, Piotr Binkowski, Johan Bjork, David Black, Tymoteusz
-Blazejczyk, Daniel Bone, Gregg Bouchard, Christopher Boumenot, Nick Bowler,
-Byron Bradley, Bryan Brady, Maarten De Braekeleer, Charlie Brej, J Briquet,
-Lane Brooks, John Brownlee, Jeff Bush, Lawrence Butcher, Tony Bybell, Iru
-Cai, Ted Campbell, Chris Candler, Lauren Carlson, Donal Casey, Alex
-Chadwick, Marcel Chang, Aliaksei Chapyzhenka, Guokai Chen, Terry Chen,
-Yi-Chung Chen, Enzo Chi, Robert A. Clark, Ryan Clarke, Allan Cochrane, John
-Coiner, Keith Colbert, Gianfranco Costamagna, Sean Cross, George Cuan, Joe
-DErrico, Lukasz Dalek, Gunter Dannoritzer, Ashutosh Das, Bernard Deadman,
-John Demme, Mike Denio, John Deroo, Philip Derrick, John Dickol, Ruben
-Diez, Danny Ding, Jacko Dirks, Ivan Djordjevic, Jonathon Donaldson, Larry
-Doolittle, Sebastian Dressler, Jonathan Drolet, Alex Duller, Jeff Dutton,
-Tomas Dzetkulic, Usuario Eda, Charles Eddleston, Chandan Egbert, Joe Eiler,
-Ahmed El-Mahmoudy, Trevor Elbourne, Mats Engstrom, Robert Farrell, Eugen
-Fekete, Fabrizio Ferrandi, Udi Finkelstein, Brian Flachs, Bill Flynn,
-Andrea Foletto, Bob Fredieu, Duane Galbi, Mostafa Gamal, Benjamin Gartner,
-Christian Gelinek, Richard E George, Peter Gerst, Glen Gibb, Michael
-Gielda, Barbara Gigerl, Shankar Giri, Dan Gisselquist, Petr Gladkikh, Sam
-Gladstone, Mariusz Glebocki, Andrew Goessling, Amir Gonnen, Chitlesh
-Goorah, Tomasz Gorochowik, Kai Gossner, Sergi Granell, Al Grant, Nathan
-Graybeal, Alexander Grobman, Graham Rushton, Xuan Guo, Driss Hafdi, Neil
-Hamilton, James Hanlon, Oyvind Harboe, Jannis Harder, Junji Hashimoto,
-Thomas Hawkins, Mitch Hayenga, Harald Heckmann, Robert Henry, Stephen
-Henry, David Hewson, Jamey Hicks, Joel Holdsworth, Andrew Holme, Hiroki
-Honda, Alex Hornung, Pierre-Henri Horrein, David Horton, Peter Horvath, Jae
-Hossell, Kuoping Hsu, Teng Huang, Steven Hugg, Alan Hunter, James
-Hutchinson, Ehab Ibrahim, Edgar E. Iglesias, Jamie Iles, Vighnesh Iyer, Ben
-Jackson, Shareef Jalloq, Marlon James, Krzysztof Jankowski, HyungKi Jeong,
-Iztok Jeras, Alexandre Joannou, James Johnson, Christophe Joly, Franck
-Jullien, James Jung, Mike Kagen, Arthur Kahlich, Kaalia Kahn, Guy-Armand
+David Addison, Tariq B. Ahmad, Nikana Anastasiadis, John David Anglin,
+Frederic Antonin, Hans Van Antwerpen, Vasu Arasanipalai, Jens Arm, Rohan
+Arshid, Gökçe Aydos, Adam Bagley, Sharad Bagri, Robert Balas, Marco
+Balboni, Matthew Ballance, Andrew Bardsley, Ilya Barkov, Matthew Barr,
+Geoff Barrett, Kaleb Barrett, Daniel Bates, Julius Baxter, Michael Berman,
+Jean Berniolles, Victor Besyakov, Narayan Bhagavatula, Moinak
+Bhattacharyya, Kritik Bhimani, David Biancolin, David Binderman, Piotr
+Binkowski, Johan Björk, David Black, Tymoteusz Blazejczyk, Scott Bleiweiss,
+David van der Bokke, Daniel Bone, Guy Bonneau, Krzysztof Boroński, Gregg
+Bouchard, Christopher Boumenot, Nick Bowler, Bryan Brady, Maarten De
+Braekeleer, Charlie Brej, J Briquet, John Brownlee, KC Buckenmaier, Jeff
+Bush, Lawrence Butcher, Tony Bybell, Iru Cai, Ted Campbell, Anthony Campos,
+Chris Candler, Lauren Carlson, Gregory Carver, Donal Casey, Sebastien Van
+Cauwenberghe, Alex Chadwick, Greg Chadwick, Marcel Chang, Aliaksei
+Chapyzhenka, Chih-Mao Chen, Guokai Chen, Terry Chen, Yi-Chung Chen, Yurii
+Cherkasov, Hennadii Chernyshchyk, Enzo Chi, Robert A. Clark, Ryan Clarke,
+Allan Cochrane, Keith Colbert, Quentin Corradi, Nassim Corteggiani,
+Gianfranco Costamagna, February Cozzocrea, Sean Cross, George Cuan, Michal
+Czyz, Joe DErrico, Jim Dai, Lukasz Dalek, Laurens van Dam, Gunter
+Dannoritzer, Ashutosh Das, Julian Daube, Bernard Deadman, Peter Debacker,
+John Demme, Mike Denio, John Deroo, Philip Derrick, Aadi Desai, John
+Dickol, Ruben Diez, Danny Ding, Jacko Dirks, Ivan Djordjevic, Brad Dobbie,
+Paul Donahue, Jonathon Donaldson, Anthony Donlon, Caleb Donovick, Larry
+Doolittle, Leendert van Doorn, Sebastian Dressler, Jonathan Drolet, Maciej
+Dudek, Alex Duller, Jeff Dutton, Tomas Dzetkulic, Usuario Eda, Charles
+Eddleston, Chandan Egbert, Joe Eiler, Ahmed El-Mahmoudy, Trevor Elbourne,
+Mats Engstrom, Robert Farrell, Julien Faucher, Olivier Faure, Eugen Fekete,
+Fabrizio Ferrandi, Udi Finkelstein, Brian Flachs, Bill Flynn, Andrea
+Foletto, Alex Forencich, Aurelien Francillon, Bob Fredieu, Manuel
+Freiberger, Mostafa Gamal, Vito Gamberini, Mostafa Garnal, Benjamin
+Gartner, Christian Gelinek, Richard E George, Peter Gerst, Glen Gibb,
+Michael Gielda, Barbara Gigerl, Nimrod Gileadi, Shankar Giri, Dan
+Gisselquist, Petr Gladkikh, Sam Gladstone, Mariusz Glebocki, Embedded Go,
+Andrew Goessling, Amir Gonnen, Chitlesh Goorah, Tomasz Gorochowik, Kai
+Gossner, Tarik Graba, Sergi Granell, Al Grant, Nathan Graybeal, Alexander
+Grobman, Qian Gu, Xuan Guo, Prabhat Gupta, Driss Hafdi, Neil Hamilton,
+James Hanlon, Tang Haojin, Øyvind Harboe, Jannis Harder, David Harris,
+Junji Hashimoto, Thomas Hawkins, Mitch Hayenga, Harald Heckmann, Robert
+Henry, Stephen Henry, Sebastian Hesselbarth, David Hewson, Jamey Hicks,
+Joel Holdsworth, Andrew Holme, Peter Holmes, Hiroki Honda, Alex Hornung,
+Pierre-Henri Horrein, David Horton, Peter Horvath, Jae Hossell, Kuoping
+Hsu, Teng Huang, Steven Hugg, Huanghuang Zhou, Alan Hunter, James
+Hutchinson, Tim Hutt, Ehab Ibrahim, Edgar E. Iglesias, Shahid Ikram, Jamie
+Iles, Vighnesh Iyer, Ben Jackson, Daniel Jacques, Shareef Jalloq, Marlon
+James, Krzysztof Jankowski, Eyck Jentzsch, HyungKi Jeong, Iztok Jeras,
+Alexandre Joannou, James Johnson, Christophe Joly, Justin Jones,
+William D. Jones, Larry Darryl Lee Jr., Franck Jullien, James Jung,
+Yoshitomo Kaneda, Mike Kagen, Arthur Kahlich, Kaalia Kahn, Guy-Armand
 Kamendje, Vasu Kandadi, Kanad Kanhere, Patricio Kaplan, Pieter Kapsenberg,
 Rafal Kapuscik, Ralf Karge, Per Karlsson, Dan Katz, Sol Katzman, Ian
-Kennedy, Michael Killough, Sun Kim, Jonathan Kimmitt, Olof Kindgren, Kevin
-Kiningham, Dan Kirkham, Aleksander Kiryk, Sobhan Klnv, Gernot Koch, Jack
-Koenig, Soon Koh, Nathan Kohagen, Steve Kolecki, Brett Koonce, Will
-Korteland, Wojciech Koszek, Varun Koyyalagunta, Arkadiusz Kozdra, Markus
-Krause, David Kravitz, Roland Kruse, Andreas Kuster, Sergey Kvachonok,
-Charles Eric LaForest, Ed Lander, Steve Lang, Stephane Laurent, Walter
-Lavino, Christian Leber, Larry Lee, Yoda Lee, Michaël Lefebvre, Igor Lesik,
-John Li, Eivind Liland, Yu Sheng Lin, Charlie Lind, Andrew Ling, Jiuyang
-Liu, Paul Liu, Derek Lockhart, Jake Longo, Geza Lore, Arthur Low, Stefan
-Ludwig, Dan Lussier, Fred Ma, Duraid Madina, Affe Mao, Julien Margetts,
+Kennedy, Ami Keren, Michael Killough, Sun Kim, Jonathan Kimmitt, Olof
+Kindgren, Kevin Kiningham, Cameron Kirk, Dan Kirkham, Aleksander Kiryk,
+Sobhan Klnv, Gernot Koch, Jack Koenig, Soon Koh, Nathan Kohagen, Steve
+Kolecki, Brett Koonce, Will Korteland, Andrei Kostovski, Wojciech Koszek,
+Varun Koyyalagunta, Arkadiusz Kozdra, Markus Krause, David Kravitz, Adam
+Krolnik, Roland Kruse, Mahesh Kumashikar, Andreas Kuster, Sergey Kvachonok,
+Charles Eric LaForest, Kevin Laeufer, Ed Lander, Steve Lang, Pierre
+Laroche, Stephane Laurent, Walter Lavino, Christian Leber, David Ledger,
+Alex Lee, Larry Lee, Yoda Lee, Michaël Lefebvre, Dag Lem, Igor Lesik, John
+Li, Kay Li, Zixi Li, Davide Libenzi, Nandor Licker, Eivind Liland, Ícaro
+Lima, Kevin Lin, Yu-Sheng Lin, Charlie Lind, Andrew Ling, Jiuyang Liu, Joey
+Liu, Paul Liu, Derek Lockhart, Jake Longo, Arthur Low, Jose Loyola, Stefan
+Ludwig, Dan Lussier, Konstantin Lübeck, Fred Ma, Liwei Ma, Duraid Madina,
+Oleh Maksymenko, Affe Mao, Julien Margetts, Chick Markley, Alexis Marquet,
 Mark Marshall, Alfonso Martinez, Unai Martinez-Corral, Adrien Le Masle,
-Yves Mathieu, Patrick Maupin, Conor McCullough, Jason McMullan, Elliot
-Mednick, David Metz, Wim Michiels, Miodrag Milanovic, Kevin Millis, Wai Sum
-Mong, Peter Monsson, Sean Moore, Dennis Muhlestein, John Murphy, Matt
-Myers, Nathan Myers, Richard Myers, Dimitris Nalbantis, Peter Nelson, Bob
-Newgard, Rachit Nigam, Paul Nitza, Yossi Nivin, Pete Nixon, Lisa Noack,
-Mark Nodine, Kuba Ober, Andreas Olofsson, Baltazar Ortiz, Aleksander Osman,
-Don Owen, James Pallister, Vassilis Papaefstathiou, Brad Parker, Morten
-Borup Petersen, Dan Petrisko, Maciej Piechotka, David Pierce, Cody
-Piersall, Michael Platzer, Dominic Plunkett, David Poole, Mike Popoloski,
-Roman Popov, Rich Porter, Stefan Post, Niranjan Prabhu, Damien Pretet, Usha
-Priyadharshini, Mark Jackson Pulver, Prateek Puri, Marshal Qiao, Nandu Raj,
-Kamil Rakoczy, Danilo Ramos, Drew Ranck, Chris Randall, Anton Rapp, Josh
-Redford, Odd Magne Reitan, Frederic Requin, Dustin Richmond, Samuel Riedel,
-Alberto Del Rio, Eric Rippey, Oleg Rodionov, Ludwig Rogiers, Paul Rolfe,
-Arjen Roodselaar, Tobias Rosenkranz, Ryszard Rozak, Huang Rui, Graham
-Rushton, Jan Egil Ruud, Denis Rystsov, John Sanguinetti, Martin Schmidt,
-Julie Schwartz, Galen Seitz, Joseph Shaker, Salman Sheikh, Hao Shi, Mike
-Shinkarovsky, Rafael Shirakawa, Jeffrey Short, Fan Shupei, Anderson Ignacio
-da Silva, Rodney Sinclair, Ameya Vikram Singh, Steven Slatter, Mladen
-Slijepcevic, Brian Small, Garrett Smith, Tim Snyder, Wilson Snyder, Maciej
+Yves Mathieu, Vladimir Matveyenko, Patrick Maupin, Stan Mayer, Conor
+McCullough, Jason McMullan, Elliot Mednick, Yuan Mei, Andy Meier,
+Rodrigo A. Melo, Benjamin Menküc, Jake Merdich, David Metz, Wim Michiels,
+Miodrag Milanović, Darryl Miles, Kevin Millis, Andrew Miloradovsky, Wai Sum
+Mong, Peter Monsson, Sean Moore, Stuart Morris, Dennis Muhlestein, John
+Murphy, Matt Myers, Nathan Myers, Richard Myers, Alex Mykyta, Dimitris
+Nalbantis, Peter Nelson, Felix Neumärker, Bob Newgard, Cong Van Nguyen,
+Rachit Nigam, Toru Niina, Paul Nitza, Yossi Nivin, Pete Nixon, Lisa Noack,
+Mark Nodine, Michael Nolan, Andrew Nolte, Joseph Nwabueze, Kuba Ober,
+Andreas Olofsson, Baltazar Ortiz, Aleksander Osman, Don Owen, Tim Paine,
+Deepa Palaniappan, James Pallister, Vassilis Papaefstathiou, Sanggyu Park,
+Brad Parker, Risto Pejašinović, Morten Borup Petersen, Dan Petrisko, Wesley
+Piard, Maciej Piechotka, David Pierce, Cody Piersall, T. Platz, Michael
+Platzer, Dominic Plunkett, David Poole, Michael Popoloski, Roman Popov,
+Aylon Chaim Porat, Oron Port, Rich Porter, Rick Porter, Stefan Post,
+Niranjan Prabhu, Damien Pretet, Harald Pretl, Bill Pringlemeir, Usha
+Priyadharshini, Mark Jackson Pulver, Prateek Puri, Jiacheng Qian, Marshal
+Qiao, Raynard Qiao, Yujia Qiao, Jasen Qin, Frank Qiu, Nandu Raj, Kamil
+Rakoczy, Danilo Ramos, Drew Ranck, Chris Randall, Anton Rapp, Josh Redford,
+Odd Magne Reitan, Frédéric Requin, Dustin Richmond, Samuel Riedel, Alberto
+Del Rio, Eric Rippey, Oleg Rodionov, Ludwig Rogiers, Paul Rolfe, Michail
+Rontionov, Arjen Roodselaar, Tobias Rosenkranz, Yernagula Roshit, Ryszard
+Rozak, Huang Rui, Graham Rushton, Jan Egil Ruud, Denis Rystsov, Pawel
+Sagan, Robert Sammelson, John Sanguinetti, Josep Sans, Luca Sasselli,
+Martin Scharrer, Martin Schmidt, Julie Schwartz, Galen Seitz, Joseph
+Shaker, Mark Shaw, Salman Sheikh, Zhou Shen, Hao Shi, James Shi, Michael
+Shinkarovsky, Rafael Shirakawa, Jeffrey Short, S Shuba, Fan Shupei, Ethan
+Sifferman, Anderson Ignacio da Silva, Rodney Sinclair, Ameya Vikram Singh,
+Sanjay Singh, Frans Skarman, Nate Slager, Steven Slatter, Mladen
+Slijepcevic, Brian Small, Garrett Smith, Gus Smith, Tim Snyder, Maciej
 Sobkowski, Stan Sokorac, Alex Solomatnikov, Flavien Solt, Wei Song, Trefor
 Southwell, Martin Stadler, Art Stamness, David Stanford, John Stevenson,
-Pete Stevenson, Patrick Stewart, Rob Stoddard, Todd Strader, John Stroebel,
-Sven Stucki, Howard Su, Emerson Suguimoto, Gene Sullivan, Qingyao Sun,
-Renga Sundararajan, Gustav Svensk, Rupert Swarbrick, Yutetsu Takatsukasa,
-Thierry Tambe, Drew Taussig, Jose Tejada, Peter Tengstrand, Wesley
-Terpstra, Rui Terra, Stefan Thiede, Gary Thomas, Ian Thompson, Kevin
-Thompson, Mike Thyer, Hans Tichelaar, Viktor Tomov, Steve Tong, Topa
-Topino, Alex Torregrosa, Michael Tresidder, David Turner, Neil Turton, Mike
-Urbach, Hans Van Antwerpen, Sebastien Van Cauwenberghe, Laurens van Dam,
-Leendert van Doorn, Cong Van Nguyen, Jan Van Winkel, Srini Vemuri, Yuri
-Victorovich, Bogdan Vukobratovic, Holger Waechtler, Philipp Wagner, Stefan
-Wallentowitz, Shawn Wang, Zhanglei Wang, Paul Wasson, Greg Waters, Thomas
-Watts, Eugene Weber, David Welch, Thomas J Whatson, Martin Whitaker, Marco
-Widmer, Leon Wildman, Daniel Wilkerson, Gerald Williams, Trevor Williams,
-Jeff Winston, Joshua Wise, Clifford Wolf, Tobias Wolfel, Johan Wouters,
-Paul Wright, Junyi Xi, Ding Xiaoliang, Jie Xu, Mandy Xu, Yinan Xu, Luke
-Yang, Amir Yazdanbakhsh, Keyi Zhang, and Xi Zhang.
+Pete Stevenson, Patrick Stewart, Rob Stoddard, Tood Strader, John Stroebel,
+Ray Strouble, Sven Stucki, Howard Su, Emerson Suguimoto, Gene Sullivan,
+Qingyao Sun, Renga Sundararajan, Kuba Sunderland-Ober, Gustav Svensk,
+Rupert Swarbrick, Jevin Sweval, Shinya T-Y, Thierry Tambe, Jesse Taube,
+Drew Taussig, Jose Tejada, Sören Tempel, Peter Tengstrand, Wesley Terpstra,
+Rui Terra, Stefan Thiede, Justin Thiel, Gary Thomas, Ian Thompson, Kevin
+Thompson, Mike Thyer, Hans Tichelaar, Tudor Timi, Viktor Tomov, Steve Tong,
+Topa Topino, Àlex Torregrosa, Topa Tota, Michael Tresidder, Lenny Truong,
+David Turner, Neil Turton, Hideto Ueno, Mike Urbach, Joel Vandergriendt,
+Srini Vemuri, Srinivasan Venkataramanan, Yuri Victorovich, Ivan Vnučec,
+Bogdan Vukobratovic, Holger Waechtler, Philipp Wagner, Johannes Walter, CY
+Wang, Chuxuan Wang, Shawn Wang, Zhanglei Wang, Greg Waters, Thomas Watts,
+Eugene Weber, John Wehle, Tianrui Wei, David Welch, Thomas J Whatson,
+Martin Whitaker, Marco Widmer, Leon Wildman, Daniel S. Wilkerson, Daniel
+Wilkerson, Gerald Williams, Trevor Williams, Don Williamson, Jan Van
+Winkel, Jeff Winston, Joshua Wise, Clifford Wolf, Johan Wouters, Paul
+Wright, Tobias Wölfel, Junyi Xi, Ding Xiaoliang, Liu Xiaoyi, Mandy Xu,
+Shanshan Xu, Yinan Xu, SU YANG, Felix Yan, Luke Yang, Amir Yazdanbakhsh,
+Chentai (Seven) Yuan, Florian Zaruba, Mat Zeno, Keyi Zhang, Xi Zhang, Yike
+Zhou, Jiamin Zhu.

 Thanks to them, and all those we've missed mentioning above, and to those
 whom have wished to remain anonymous.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 *********
@@ -8,7 +8,7 @@ Copyright
 The latest version of Verilator is available from `https://verilator.org
 <https://verilator.org>`_.

-Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 can redistribute it and/or modify the Verilator internals under the terms
 of either the GNU Lesser General Public License Version 3 or the Perl
 Artistic License Version 2.0.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 Deprecations
@@ -8,36 +8,13 @@ The following deprecated items are scheduled for future removal:

 C++11 compiler support
  Verilator currently requires a C++20 or newer compiler for timing, and a
-  C++11 or newer compiler for no-timing.
-
-  Verilator will require C++14 or newer compilers for both compiling
-  Verilator and compiling Verilated models with --no-timing no sooner than
-  January 2023.
+  C++14 or newer compiler for both compiling Verilator and compiling
+  Verilated models with --no-timing.

  Verilator will require C++20 or newer compilers for both compiling
  Verilator and compiling all Verilated models no sooner than January 2025.

-Verilated_heavy.h
-  The legacy "verilated_heavy.h" include was replaced with just including
-  "verilated.h". Verilated_heavy.h is planned for removal no sooner than
-  July 2022.
-
-Option `--cdc`
-  The experimental `--cdc` option is believed to be generally unused and is
-  planned for removal no sooner than January 2023.
-
-Option `-O<letter>`
-  The debug `-O<letter>` options have been replaced with
-  `-fno-<optimization>` debug options to match GCC. The old options are
-  planned for removal no sooner than June 2023.
-
-Option `--prof-threads`
-  The `--prof-threads` option has been superseded by the `--prof-exec` and
-  `--prof-pgo` options and is planned for removal no sooner than April 2023.
-
-Verilated model options `+verilator+prof+threads+*`
-  The `+verilator+prof+threads+start`, `+verilator+prof+threads+window` and
-  `+verilator+prof+threads+file` options have been superseded by the
-  `+verilator+prof+exec+start`, `+verilator+prof+exec+window` and
-  `+verilator+prof+exec+file` options respectively and are planned for removal
-  no sooner than April 2023.
+32-bit compiler support
+  Verilator currently regresses both 64-bit and 32-bit pointer modes (GCC's
+  `-m64` and `-m32`).  Support for 32-bit `-m32` mode will be deprecated no
+  sooner than January 2024.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 Environment
@@ -29,7 +29,7 @@ associated programs.
 .. option:: OBJCACHE

   Optionally specifies a caching or distribution program to place in front
-   of all runs of the C++ compiler.  For example, "ccache".  If using
+   of all runs of the C++ compiler.  For example, "ccache" or "sccache".  If using
   :command:`distcc` or :command:`icecc`/:command:`icecream`, they would
   generally be run under :command:`ccache`; see the documentation for
   those programs.  If OBJCACHE is not set, and at configure time ccache
@@ -91,9 +91,72 @@ associated programs.

 .. option:: VERILATOR_ROOT

-   Specifies the directory containing the distribution kit.  This is used
-   to find the executable, Perl library, and include files.  If not
-   specified, it will come from a default optionally specified at configure
-   time (before Verilator was compiled).  It should not be specified if
-   using a pre-compiled Verilator package as the hard-coded value should be
-   correct.  See :ref:`Installation`.
+   The ``VERILATOR_ROOT`` environment variable is used in several places:
+
+   * At ``./configure`` time: If set, it is embedded into the binary, and
+     at runtime if ``VERILATOR_ROOT`` is not set, the embedded value is
+     used for the runtime default.
+
+   * When ``verilator`` is run: If ``VERILATOR_ROOT`` is set it will be
+     used to find the ``verilator_bin`` executable (this is the actual
+     Verilator binary; ``verilator`` is a Perl wrapper). If not set, the
+     ``verilator`` script uses other methods to find ``verilator_bin``
+     (looking in the same directory and falling back to ``$PATH``).
+
+   * When ``make`` is run on the Makefile generated by ``verilator``: The
+     value of ``VERILATOR_ROOT`` (falling back to the value embedded in the
+     binary if not set) is used to find the include files
+     (``include/verilated.mk``).
+
+   If you are using a pre-compiled Verilator package, you should not need
+   to set ``VERILATOR_ROOT`` - the value embedded in the binary should be
+   correct.  In fact this option *does not work* with Verilator packages
+   that have been installed with ``make install``. If a Verilator package
+   has been installed using ``./configure --prefix=/some/path && make
+   install`` and then moved to another location, you cannot use
+   ``VERILATOR_ROOT`` to point to the new version.
+
+   See :ref:`Installation` for more details.
+
+
+Make Variables
+==============
+
+This section describes the make variables used by Verilator. These may be
+set by passing them to make e.g. ``make CXX=my-gcc ...``.
+
+.. option:: AR
+
+   Optionally overrides the default :command:`ar` (archive) binary used by
+   the Verilated makefiles. If AR is not set, the version found at
+   configure time is used.
+
+.. option:: CXX
+
+   Optionally overrides the default compiler binary used by the Verilated
+   makefiles. If CXX is not set, the version found at configure time is
+   used.  Note the default flags passed to the compiler are determined at
+   configuration time, so changing the CXX compiler version using this
+   variable, as opposed to passing it at configuration time, may not give
+   desired results.
+
+.. option:: LINK
+
+   Optionally overrides the default linker binary used by the Verilated
+   makefiles. If LINK is not set, the version found at configure time is
+   used.  Note the default flags passed to the linker are determined at
+   configuration time, so changing the LINK version using this variable, as
+   opposed to passing it at configuration time, may not give desired
+   results.
+
+.. option:: PERL
+
+   Optionally overrides the default :command:`perl` binary used by the
+   Verilated makefiles. If PERL is not set, the version found at configure
+   time, and compiled into the Verilator binary, is used.
+
+.. option:: PYTHON3
+
+   Optionally overrides the default :command:`python3` binary used by the
+   Verilated makefiles. If PYTHON3 is not set, the version found at
+   configure time is used.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Example Create-Binary Execution:
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Example C++ Execution:
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 First you need Verilator installed, see :ref:`Installation`.  In brief, if
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Examples in the Distribution:
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Example SystemC Execution:
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Examples:
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Simulation Runtime Arguments:
@@ -64,15 +64,18 @@ Summary:

 .. option:: +verilator+prof+threads+file+<filename>

-   Deprecated. Alias for :vlopt:`+verilator+prof+exec+file+\<filename\>`
+   Removed in 5.020. Was an alias for
+   :vlopt:`+verilator+prof+exec+file+\<filename\>`

 .. option:: +verilator+prof+threads+start+<value>

-   Deprecated. Alias for :vlopt:`+verilator+prof+exec+start+\<value\>`
+   Removed in 5.020. Was an alias for
+   :vlopt:`+verilator+prof+exec+start+\<value\>`

 .. option:: +verilator+prof+threads+window+<value>

-   Deprecated. Alias for :vlopt:`+verilator+prof+exec+window+\<value\>`
+   Removed in 5.020. Was an alias for
+   :vlopt:`+verilator+prof+exec+window+\<value\>`

 .. option:: +verilator+prof+vlt+file+<filename>

@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 verilator Arguments
@@ -125,9 +125,12 @@ Summary:

   After generating the SystemC/C++ code, Verilator will invoke the
   toolchain to build the model library (and executable when :vlopt:`--exe`
-   is also used). Verilator manages the build itself, and for this --build
+   is also used).  Verilator manages the build itself, and for this --build
   requires GNU Make to be available on the platform.

+   :vlopt:`--build` cannot be specified when using :vlopt:`-E`,
+   :vlopt:`--dpi-hdr-only`, :vlopt:`--lint-only`, or :vlopt:`--xml-only`.
+
 .. option:: --build-dep-bin <filename>

   Rarely needed.  When a dependency (.d) file is created, this filename
@@ -155,18 +158,6 @@ Summary:
   Specify C++ without SystemC output mode; see also the :vlopt:`--sc`
   option.

-.. option:: --cdc
-
-   Permanently experimental.  Perform some clock domain crossing checks and
-   issue related warnings (CDCRSTLOGIC) and then exit; if warnings other
-   than CDC warnings are needed, make a second run with
-   :vlopt:`--lint-only`.  Additional warning information is also written to
-   the file :file:`<prefix>__cdc.txt`.
-
-   Currently only checks some items that other CDC tools missed; if you are
-   interested in adding more traditional CDC checks, please contact the
-   authors.
-
 .. option:: -CFLAGS <flags>

   Add specified C compiler argument to the generated makefiles. For
@@ -434,7 +425,7 @@ Summary:
   Preprocess the source code, but do not compile, similar to C++
   preprocessing using :command:`gcc -E`.  Output is written to standard
   out.  Beware of enabling debugging messages, as they will also go to
-   standard out.
+   standard out. See :vlopt:`--no-std`, which is implied by this.

   See also :vlopt:`--dump-defines`, :vlopt:`-P`, and
   :vlopt:`--pp-comments` options.
@@ -707,9 +698,10 @@ Summary:
 .. option:: -j [<value>]

   Specify the level of parallelism for :vlopt:`--build` if
-   :vlopt:`--build-jobs` isn't provided. If zero, uses the number of threads
-   in the current hardware. Otherwise, the <value> must be a positive
-   integer specifying the maximum number of parallel build jobs.
+   :vlopt:`--build-jobs` isn't provided, and the internal compilation steps
+   of Verilator if :vlopt:`--verilate-jobs` isn't provided. If zero, uses
+   the number of threads in the current hardware. Otherwise, must be a
+   positive integer specifying the maximum number of parallel build jobs.

 .. option:: --l2-name <value>

@@ -819,6 +811,13 @@ Summary:

   See also :vlopt:`--binary`.

+.. option:: --main-top-name <string>
+
+   Specify the name passed to the Verilated model being constructed, in the
+   generated C++ main() function.
+
+   If the string ``"-"`` is used, no top level scope is added.
+
 .. option:: --max-num-width <value>

   Set the maximum number literal width (e.g., in 1024'd22 this
@@ -1037,7 +1036,7 @@ Summary:

 .. option:: --prof-threads

-   Deprecated. Same as --prof-exec and --prof-pgo together.
+   Removed in 5.020. Was an alias for --prof-exec and --prof-pgo together.

 .. option:: --protect-ids

@@ -1112,6 +1111,18 @@ Summary:
   marking only those signals that need public_flat_rw is typically
   significantly better performing.

+.. option:: --public-depth <level>
+
+   Enables public as with :vlopt:`--public-flat-rw`, but only to the specified depth of modules.
+   It operates at the module maximum level, so if a module's cells are A.B.X and A.X, the
+   a --public-depth 3 must be used to make module X public, and both A.B.X and A.X will be public.
+
+.. option:: --public-params
+
+   Declares all parameters public as if they had
+   :code:`/*verilator public_flat_rd*/`
+   metacomments.
+
 .. option:: -pvalue+<name>=<value>

   Overwrites the given parameter(s) of the top-level module. See
@@ -1201,6 +1212,10 @@ Summary:
   by size (plain :vlopt:`--stats` just gives a count).  See
   :vlopt:`--stats`, which is implied by this.

+.. option:: --no-std
+
+   Prevents parsing standard library.
+
 .. option:: --structs-packed

   Converts all unpacked structures to packed structures, and issues an
@@ -1302,6 +1317,7 @@ Summary:
   This is not needed with standard designs with only one top.  See also
   :option:`MULTITOP` warning.

+
 .. option:: --trace

   Adds waveform tracing code to the model using VCD format. This overrides
@@ -1347,13 +1363,13 @@ Summary:
   Enable FST waveform tracing in the model. This overrides
   :vlopt:`--trace`.  See also :vlopt:`--trace-threads` option.

-.. option:: --trace-max-array *depth*
+.. option:: --trace-max-array <depth>

   Rarely needed.  Specify the maximum array depth of a signal that may be
   traced.  Defaults to 32, as tracing large arrays may greatly slow traced
   simulations.

-.. option:: --trace-max-width *width*
+.. option:: --trace-max-width <width>

   Rarely needed.  Specify the maximum bit width of a signal that may be
   traced.  Defaults to 256, as tracing large vectors may greatly slow
@@ -1370,7 +1386,7 @@ Summary:
   format constraints, this may result in significantly slower trace times
   and larger trace files.

-.. option:: --trace-threads *threads*
+.. option:: --trace-threads <threads>

   Enable waveform tracing using separate threads. This is typically faster
   in simulation runtime but uses more total compute. This option only
@@ -1380,6 +1396,17 @@ Summary:
   This option is accepted, but has absolutely no effect with
   :vlopt:`--trace`, which respects :vlopt:`--threads` instead.

+.. option:: --no-trace-top
+
+   Disables tracing for the input and output signals in the top wrapper which
+   Verilator adds to the design. The signals are still traced in the original
+   verilog top modules.
+
+   When combined with :option:`--main-top-name` set to "-" or when the name of
+   the top module is set to "" in its constructor, the generated trace file
+   will have the verilog top module as its root, rather than another module
+   added by Verilator.
+
 .. option:: --trace-underscore

   Enable tracing of signals or modules that start with an
@@ -1400,12 +1427,12 @@ Summary:
   Rarely needed.  Specifies the maximum number of loop iterations that may be
   unrolled.  See also :option:`BLKLOOPINIT` warning.

-.. option:: --unroll-stmts *statements*
+.. option:: --unroll-stmts <statements>

   Rarely needed.  Specifies the maximum number of statements in a loop for
   that loop to be unrolled. See also :option:`BLKLOOPINIT` warning.

-.. option:: --unused-regexp *regexp*
+.. option:: --unused-regexp <regexp>

   Rarely needed.  Specifies a simple regexp with \* and ? that, if a signal
   name matches, will suppress the :option:`UNUSED` warning.  Defaults to
@@ -1417,7 +1444,7 @@ Summary:
   into Verilator.  (Similar to :command:`perl -V`.)  See also
   :vlopt:`--getenv` option.

-.. option:: -v *filename*
+.. option:: -v <filename>

   Read the filename as a Verilog library.  Any modules in the file may be
   used to resolve instances in the top-level module, otherwise, they are
@@ -1429,6 +1456,15 @@ Summary:
   execute only the build. This can be useful for rebuilding the Verilated code
   produced by a previous invocation of Verilator.

+.. option:: --verilate-jobs [<value>]
+
+   Specify the level of parallelism for the internal compilation steps of
+   Verilator. If zero, uses the number of threads in the current hardware.
+   Otherwise, must be a positive integer specifying the maximum number of
+   parallel build jobs.
+
+   See also :vlopt:`-j`.
+
 .. option:: +verilog1995ext+<ext>

   Synonym for :vlopt:`+1364-1995ext+\<ext\>`.
@@ -1445,7 +1481,7 @@ Summary:

   Enable the use of VPI and linking against the :file:`verilated_vpi.cpp` files.

-.. option:: --waiver-output *filename*
+.. option:: --waiver-output <filename>

   Generate a waiver file that contains all waiver statements to suppress
   the warnings emitted during this Verilator run. This, in particular, is
@@ -1453,8 +1489,11 @@ Summary:
   them systematically.

   The generated file is in the Verilator Configuration format, see
-   :ref:`Configuration Files`, and can directly be consumed by
-   Verilator. The standard file extension is ".vlt".
+   :ref:`Configuration Files`. The standard file extension is ".vlt".
+   These files can directly be consumed by Verilator, typically by placing
+   the filename as part of the Verilator command line options. Waiver files
+   need to be listed on the command line before listing the files they are
+   waiving.

 .. option:: -Wall

@@ -1499,13 +1538,15 @@ Summary:

 .. option:: -Wno-lint

-   Disable all lint-related warning messages, and all style warnings.  This is
-   equivalent to ``-Wno-ALWCOMBORDER -Wno-BSSPACE -Wno-CASEINCOMPLETE
-   -Wno-CASEOVERLAP -Wno-CASEX -Wno-CASTCONST -Wno-CASEWITHX -Wno-CMPCONST -Wno-COLONPLUS
-   -Wno-IMPLICIT -Wno-IMPLICITSTATIC -Wno-LITENDIAN -Wno-PINCONNECTEMPTY
-   -Wno-PINMISSING -Wno-SYNCASYNCNET -Wno-UNDRIVEN -Wno-UNSIGNED
-   -Wno-UNUSEDGENVAR -Wno-UNUSEDPARAM -Wno-UNUSEDSIGNAL
-   -Wno-WIDTH`` plus the list shown for Wno-style.
+   Disable all lint-related warning messages, and all style warnings.  This
+   is equivalent to ``-Wno-ALWCOMBORDER`` ``-Wno-ASCRANGE``
+   ``-Wno-BSSPACE`` ``-Wno-CASEINCOMPLETE`` ``-Wno-CASEOVERLAP``
+   ``-Wno-CASEX`` ``-Wno-CASTCONST`` ``-Wno-CASEWITHX`` ``-Wno-CMPCONST``
+   ``-Wno-COLONPLUS`` ``-Wno-IMPLICIT`` ``-Wno-IMPLICITSTATIC``
+   ``-Wno-PINCONNECTEMPTY`` ``-Wno-PINMISSING`` ``-Wno-STATICVAR``
+   ``-Wno-SYNCASYNCNET`` ``-Wno-UNDRIVEN`` ``-Wno-UNSIGNED``
+   ``-Wno-UNUSEDGENVAR`` ``-Wno-UNUSEDPARAM`` ``-Wno-UNUSEDSIGNAL``
+   ``-Wno-WIDTH``, plus the list shown for :vlopt:`-Wno-style`.

   It is strongly recommended that you clean up your code rather than using this
   option; it is only intended to be used when running test-cases of code
@@ -1513,12 +1554,13 @@ Summary:

 .. option:: -Wno-style

-   Disable all code style related warning messages (note that by default, they are
-   already disabled).  This is equivalent to ``-Wno-DECLFILENAME -Wno-DEFPARAM
-   -Wno-EOFNEWLINE -Wno-IMPORTSTAR -Wno-INCABSPATH -Wno-PINCONNECTEMPTY
-   -Wno-PINNOCONNECT -Wno-SYNCASYNCNET -Wno-UNDRIVEN
-   -Wno-UNUSEDGENVAR -Wno-UNUSEDPARAM -Wno-UNUSEDSIGNAL
-   -Wno-VARHIDDEN``.
+   Disable all code style related warning messages (note that by default,
+   they are already disabled).  This is equivalent to ``-Wno-DECLFILENAME``
+   ``-Wno-DEFPARAM`` ``-Wno-EOFNEWLINE`` ``-Wno-GENUNNAMED``
+   ``-Wno-IMPORTSTAR`` ``-Wno-INCABSPATH`` ``-Wno-PINCONNECTEMPTY``
+   ``-Wno-PINNOCONNECT`` ``-Wno-SYNCASYNCNET`` ``-Wno-UNDRIVEN``
+   ``-Wno-UNUSEDGENVAR`` ``-Wno-UNUSEDPARAM`` ``-Wno-UNUSEDSIGNAL``
+   ``-Wno-VARHIDDEN``.

 .. option:: -Wpedantic

@@ -1528,25 +1570,33 @@ Summary:
   :command:`gcc -Wpedantic`.  Rarely used, and intended only for strict
   compliance tests.

+   This option changes :option:`ASSIGNIN` from an error to a warning.
+
 .. option:: -Wwarn-<message>

   Enables the specified warning message.

 .. option:: -Wwarn-lint

-   Enable all lint-related warning messages (note that by default, they are already
-   enabled), but do not affect style messages.  This is equivalent to
-   ``-Wwarn-ALWCOMBORDER -Wwarn-BSSPACE -Wwarn-CASEINCOMPLETE
-   -Wwarn-CASEOVERLAP -Wwarn-CASEX -Wwarn-CASTCONST -Wwarn-CASEWITHX -Wwarn-CMPCONST
-   -Wwarn-COLONPLUS -Wwarn-IMPLICIT -Wwarn-LITENDIAN
-   -Wwarn-PINMISSING -Wwarn-REALCVT -Wwarn-UNSIGNED -Wwarn-WIDTH``.
+   Enable all lint-related warning messages (note that by default, they are
+   already enabled), but do not affect style messages.  This is equivalent
+   to ``-Wwarn-ALWCOMBORDER`` ``-Wwarn-ASCRANGE`` ``-Wwarn-BSSPACE``
+   ``-Wwarn-CASEINCOMPLETE`` ``-Wwarn-CASEOVERLAP`` ``-Wwarn-CASEWITHX``
+   ``-Wwarn-CASEX`` ``-Wwarn-CASTCONST`` ``-Wwarn-CMPCONST``
+   ``-Wwarn-COLONPLUS`` ``-Wwarn-IMPLICIT`` ``-Wwarn-IMPLICITSTATIC``
+   ``-Wwarn-LATCH`` ``-Wwarn-MISINDENT`` ``-Wwarn-NEWERSTD``
+   ``-Wwarn-PINMISSING`` ``-Wwarn-REALCVT`` ``-Wwarn-STATICVAR``
+   ``-Wwarn-UNSIGNED`` ``-Wwarn-WIDTHTRUNC`` ``-Wwarn-WIDTHEXPAND``
+   ``-Wwarn-WIDTHXZEXPAND``.

 .. option:: -Wwarn-style

   Enable all code style-related warning messages.  This is equivalent to
-   ``-Wwarn ASSIGNDLY -Wwarn-DECLFILENAME -Wwarn-DEFPARAM -Wwarn-EOFNEWLINE
-   -Wwarn-INCABSPATH -Wwarn-PINNOCONNECT -Wwarn-SYNCASYNCNET -Wwarn-UNDRIVEN
-   -Wwarn-UNUSEDGENVAR -Wwarn-UNUSEDPARAM -Wwarn-UNUSEDSIGNAL -Wwarn-VARHIDDEN``.
+   ``-Wwarn-ASSIGNDLY`` ``-Wwarn-DECLFILENAME`` ``-Wwarn-DEFPARAM``
+   ``-Wwarn-EOFNEWLINE`` ``-Wwarn-GENUNNAMED`` ``-Wwarn-INCABSPATH``
+   ``-Wwarn-PINNOCONNECT`` ``-Wwarn-SYNCASYNCNET`` ``-Wwarn-UNDRIVEN``
+   ``-Wwarn-UNUSEDGENVAR`` ``-Wwarn-UNUSEDPARAM`` ``-Wwarn-UNUSEDSIGNAL``
+   ``-Wwarn-VARHIDDEN``.

 .. option:: --x-assign 0

@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 verilator_coverage
@@ -6,10 +6,10 @@ verilator_coverage

 Verilator_coverage processes Verilated model-generated coverage reports.

-With --annotate, it reads the specified coverage data file and generates
-annotated source code with coverage metrics annotated.  If multiple
-coverage points exist on the same source code line, additional lines will
-be inserted to report the additional points.
+With `--annotate`, it reads the specified coverage data file and generates
+annotated source code with coverage metrics annotated.  With
+`--annotate-points` the coverage points corresponding to each line are also
+shown.

 Additional Verilog-XL-style standard arguments specify the search paths
 necessary to find the source code on which the coverage analysis was
@@ -58,17 +58,65 @@ to read multiple inputs.  If no data file is specified, by default,
 Specifies the directory name to which source files with annotated coverage
 data should be written.

+Converting from the Verilator coverage data format to the info format is
+lossy; the info will have all forms of coverage merged line coverage, and
+if there are multiple coverage points on a single line they will merge.
+The minimum coverage across all merged points will be used to report
+coverage of the line.
+
+Coverage data is annotated at the beginning of the line and is formatted
+as a special character followed by the number of coverage hits. The special
+characters " ,%,+,-" indicate summary of the coverage, and allow use of grep
+to filter the report.
+
+* " " (whitespace) indicates that all points on the line are above the coverage limit.
+* "%" indicates at least one point on the line was below the coverage limit.
+* "+" coverage point was at or above the limit. Only used with :option:`--annotate-points`.
+* "-" coverage point was below the limit.  Only used with :option:`--annotate-points`.
+
+.. code-block::
+
+   100000  input logic a;   // Begins with whitespace, because
+                            // number of hits (100000) is above the limit.
+  +100000  point: comment=a // Begins with +, because
+                            // number of hits (100000) is above the limit.
+  %000000  input logic b;   // Begins with %, because
+                            // number of hits (0) is below the limit.
+  -000000  point: comment=b // Begins with -, because
+                            // number of hits (0) is below the limit.
+
 .. option:: --annotate-all

 Specifies all files should be shown.  By default, only those source files
 with low coverage are written to the output directory.

+This option should be used together with :option:`--annotate`.
+
 .. option:: --annotate-min <count>

-Specifies if the coverage point does not include the count number of
-coverage hits, then the coverage point will be considered above the
-threshold, and the coverage report will put a "%" to indicate the coverage
-is insufficient.  Defaults to 10.
+Specifies the threshold (<count>) below which coverage point is considered
+sufficient. If the threshold is not exceeded, then the annotation will begin
+with a "%" symbol to indicate the coverage is insufficient.
+
+The <count> threshold defaults to 10.
+
+This option should be used together with :option:`--annotate`.
+
+
+.. option:: --annotate-points
+
+Specifies all coverage points should be shown after each line of text.  By
+default, only source lines are shown.
+
+.. code-block::
+
+  100000  input logic a, b, c;
+ +100000 point: comment=a // These lines are only shown
+ +200000 point: comment=b // with option --annotate-points
+ +300000 point: comment=c // enabled.
+
+
+This option should be used together with :option:`--annotate`.

 .. option:: --help

@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 verilator_gantt
@@ -30,15 +30,13 @@ Gantt Chart VCD Signals
 In waveforms, there are the following signals. In GTKWave, use "decimal"
 data format to remove the leading zeros and make the traces easier to read.

-evals
-  Increments each time when eval_step was measured to be active.  This
-  allow visualization of how much time eval_step was active.
+trace/section
+  Shows the name of the current top of the execution section stack.
+  Set GTKWave data format to "ASCII".

-eval_loop
-  Increments each time when the evaluation loop within eval_step was
-  measured to be active.  For best performance, there is only a single
-  evaluation loop within each eval_step call; that is, the eval_loop
-  waveform looks identical to the evals waveform.
+trace/depth
+  Shows the depth of the execution section stack.
+  Set GTKWave data format to "Analog".

 measured_parallelism
  The number of mtasks active at this time, for best performance, this will
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 verilator_profcfunc
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 *********************************
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 *******************
@@ -17,13 +17,21 @@ or "`ifdef`"'s may break other tools.

 .. option:: `__LINE__

-   The :option:`\`__LINE__` define expands to the current filename as a
-   string, like C++'s __LINE__.  This Verilator feature added in 2006 was
-   incorporated into IEEE 1800-2009.
+   The :option:`\`__LINE__` define expands to the current line number like
+   C++'s __LINE__.  This Verilator feature added in 2006 was incorporated
+   into IEEE 1800-2009.

 .. option:: `error [string]

-   This will report an error when encountered, like C++'s #error.
+   This will report an error when the preprocessor emits it, similar to
+   C++'s #error.
+
+.. option:: `line
+
+   As a special case `\`line \`__LINE__ "filename"` allows setting the
+   filename, without changing the line number. This is used for some
+   internal tests, so that debugging can leave the line numbers correctly
+   referring to the test file's line numbers.

 .. option:: """ [string] """

@@ -446,6 +454,33 @@ or "`ifdef`"'s may break other tools.

   Same as :option:`public_flat_rw` configuration file option.

+.. option:: /*verilator&32;public_[|flat|flat_rd|flat_rw]_on [@(<edge_list>)]*/ (as scope)
+
+   Used to wrap multiple signals and parameters with the respective public attribute.
+   See attribute above for their respective behavior. Cannot be nested. e.g:
+
+      .. code-block:: sv
+
+         /*verilator public_flat_rw_on*/
+         logic clk;
+         logic rst;
+         parameter width = 8;
+         /* verilator public_off*/
+         logic data;
+
+   Is equivalent to:
+
+      .. code-block:: sv
+
+         logic clk /*verilator public_flat_rw*/;
+         logic rst /*verilator public_flat_rw*/;
+         parameter width /*verilator public_flat_rw*/ = 8;
+         logic data;
+
+.. option:: /*verilator&32;public_off*/
+
+   Terminates the previous `/*verilator public*_on*/` directive; see above.
+
 .. option:: /*verilator&32;public_module*/

   Used after a module statement to indicate the module should not be
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 ******************************
@@ -220,12 +220,6 @@ This is done for you if you are using the Verilator :vlopt:`--binary` or
 You can call :code:`->trace()` on multiple Verilated objects with the same
 trace file if you want all data to land in the same output file.

-When using SystemC 2.3, the SystemC library must have been built with the
-experimental simulation phase callback-based tracing disabled. This is
-disabled by default when building SystemC with its configure based build
-system, but when building SystemC with CMake, you must pass
-``-DENABLE_PHASE_CALLBACKS_TRACING=OFF`` to disable this feature.
-

 How do I generate FST waveforms (traces) in C++ or SystemC?
 """""""""""""""""""""""""""""""""""""""""""""""""""""""""""
@@ -356,8 +350,8 @@ Why do I get "undefined reference to \`VL_RAND_RESET_I' or \`Verilated::...'"?
 You need to link your compiled Verilated code against the
 :code:`verilated.cpp` file found in the include directory of the Verilator
 kit.  This is one target in the ``$(VK_GLOBAL_OBJS)`` make variable, which
-should be part of your Makefile's link rule.  If you use :vlopt:`--exe`,
-this is done for you.
+should be part of your Makefile's link rule.  If you use :vlopt:`--exe` or
+:vlopt:`--binary`, this is done for you.


 Is the PLI supported?
@@ -401,7 +395,7 @@ How do I get faster build times?
 * Use a recent compiler.  Newer compilers tend to be faster.

 * Compile in parallel on many machines and use caching; see the web for the
-  ccache, distcc, and icecream packages. ccache will skip GCC runs between
+  ccache, sccache, distcc, or icecream packages. ccache will skip GCC runs between
  identical source builds, even across different users.  If ccache was
  installed when Verilator was built, it is used, or see OBJCACHE
  environment variable to override this. Also see the
@@ -486,7 +480,7 @@ From the sc_main.cpp file, you'd then:

      #include "Vour.h"
      #include "Vour_our.h"
-      cout << "clock is " << top->our->clk << endl;
+      std::cout << "clock is " << top->our->clk << std::endl;


 In this example, clk is a bool you can read or set as any other variable.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 *****
@@ -50,9 +50,9 @@ For --cc/--sc, it creates:
     - Make include file with class names (from --make gmake)
   * - *{prefix}*\ _hier.mk
     - Make file for hierarchy blocks (from --make gmake)
-   * - *{prefix}*\ _hierMkArgs.f
+   * - *{prefix}*\ __hierMkArgs.f
     - Arguments for hierarchical Verilation (from --make gmake)
-   * - *{prefix}*\ _hierCMakeArgs.f
+   * - *{prefix}*\ __hierCMakeArgs.f
     - Arguments for hierarchical Verilation (from --make cmake)
   * - *{prefix}*\ .h
     - Model header
@@ -126,8 +126,10 @@ In specific debug and other modes, it also creates:
     - Debugging graph files (from --debug)
   * - *{prefix}{misc}*\ .tree
     - Debugging files (from --debug)
-   * - {mod_prefix}_{each_verilog_base_filename}*\ .vpp
-     - Pre-processed verilog (from --debug)
+   * - *{prefix}*\ __inputs\ .vpp
+     - Pre-processed verilog for all files (from --debug)
+   * - *{prefix}*\ _ *{each_verilog_base_filename}*\ .vpp
+     - Pre-processed verilog for each file (from --debug)

 After running Make, the C++ compiler may produce the following:

@@ -143,6 +145,10 @@ After running Make, the C++ compiler may produce the following:
     - Intermediate objects
   * - *{prefix}*\
     - Final executable (from --exe)
+   * - lib\ *{prefix}*\ .a
+     - Final archive (default lib mode)
+   * - libverilated.a
+     - Runtime for verilated model (default lib mode)
   * - *{prefix}*\ __ALL.a
     - Library of all Verilated objects
   * - *{prefix}*\ __ALL.cpp
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 ######################
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _CMakeInstallation:
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Installation:
@@ -39,7 +39,7 @@ In brief, to install from git:
 ::

   # Prerequisites:
-   #sudo apt-get install git perl python3 make autoconf g++ flex bison ccache
+   #sudo apt-get install git help2man perl python3 make autoconf g++ flex bison ccache
   #sudo apt-get install libgoogle-perftools-dev numactl perl-doc
   #sudo apt-get install libfl2  # Ubuntu only (ignore if gives error)
   #sudo apt-get install libfl-dev  # Ubuntu only (ignore if gives error)
@@ -104,6 +104,7 @@ for good performance:
 ::

   sudo apt-get install ccache  # If present at build, needed for run
+   sudo apt-get install mold  # If present at build, needed for run
   sudo apt-get install libgoogle-perftools-dev numactl

 The following is optional but is recommended for nicely rendered command line
@@ -124,9 +125,9 @@ Those developing Verilator itself may also want these (see internals.rst):

 ::

-   sudo apt-get install gdb graphviz cmake clang clang-format-14 gprof lcov
-   sudo apt-get install libclang-dev yapf3
-   sudo pip3 install clang sphinx sphinx_rtd_theme sphinxcontrib-spelling breathe
+   sudo apt-get install clang clang-format-14 cmake gdb gprof graphviz lcov
+   sudo apt-get install libclang-dev yapf3 bear
+   sudo pip3 install clang sphinx sphinx_rtd_theme sphinxcontrib-spelling breathe ruff
   cpan install Pod::Perldoc
   cpan install Parallel::Forker

@@ -201,8 +202,9 @@ These are the installation options:
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

 Our personal favorite is to always run Verilator in-place from its Git
-directory. This allows the easiest experimentation and upgrading, and
-allows many versions of Verilator to co-exist on a system.
+directory (don't run ``make install``). This allows the easiest
+experimentation and upgrading, and allows many versions of Verilator to
+co-exist on a system.

 ::

@@ -211,17 +213,18 @@ allows many versions of Verilator to co-exist on a system.
   ./configure
   # Running will use files from $VERILATOR_ROOT, so no install needed

-Note after installing (below steps), a calling program or shell must set
-the environment variable :option:`VERILATOR_ROOT` to point to this Git
-directory, then execute ``$VERILATOR_ROOT/bin/verilator``, which will find
-the path to all needed files.
+Note after installing (see `Installation`_), a calling program or shell
+must set the environment variable :option:`VERILATOR_ROOT` to point to this
+Git directory, then execute ``$VERILATOR_ROOT/bin/verilator``, which will
+find the path to all needed files.


-2. Install into a specific location
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+2. Install into a Specific Prefix
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

-You may eventually be installing onto a project/company-wide "CAD" tools
-disk that may support multiple versions of every tool. Tell configure the
+You may be an OS package maintainer building a Verilator package, or you
+may eventually be installing onto a project/company-wide "CAD" tools disk
+that may support multiple versions of every tool. Tell configure the
 eventual destination directory name.  We recommend that the destination
 location include the Verilator version name:

@@ -232,7 +235,8 @@ location include the Verilator version name:
   # For the tarball, use the version number instead of git describe
   ./configure --prefix /CAD_DISK/verilator/`git describe | sed "s/verilator_//"`

-Note after installing (below steps), if you use `modulecmd
+Note after installing (see `Installation`_), you need to add the path to
+the ``bin`` directory to your ``PATH``. Or, if you use `modulecmd
 <http://modules.sourceforge.net/>`__, you'll want a module file like the
 following:

@@ -245,23 +249,7 @@ following:
   prepend-path PKG_CONFIG_PATH $install_root/share/pkgconfig


-3. Install into a Specific Path
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-You may eventually install Verilator into a specific installation prefix,
-as most GNU tools support:
-
-::
-
-   unset VERILATOR_ROOT      # if your shell is bash
-   unsetenv VERILATOR_ROOT   # if your shell is csh
-   ./configure --prefix /opt/verilator-VERSION
-
-Then after installing (below steps), you will need to add
-``/opt/verilator-VERSION/bin`` to your ``$PATH`` environment variable.
-
-
-4. Install System Globally
+3. Install System Globally
 ^^^^^^^^^^^^^^^^^^^^^^^^^^

 The final option is to eventually install Verilator globally, using
@@ -273,8 +261,8 @@ configure's default system paths:
   unsetenv VERILATOR_ROOT   # if your shell is csh
   ./configure

-Then after installing (below), the binaries should be in a location
-already in your ``$PATH`` environment variable.
+Then after installing (see `Installation`_), the binaries should be in a
+location already in your ``$PATH`` environment variable.


 Configure
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 ***************
@@ -237,11 +237,9 @@ optimizations will be disabled around the latch.
 Structures and Unions
 ---------------------

-Presently Verilator only supports packed structs and packed unions.  Rand
-and randc tags on members are ignored.  All structures and unions are
-represented as a single vector, which means that generating one member of a
-structure from blocking, and another from non-blocking assignments is
-unsupported.
+All structures and unions are scheduled together, which means that
+generating one member of a structure from blocking, and another from
+non-blocking assignments is unsupported.


 .. _Unknown States:
@@ -320,20 +318,6 @@ there will be considered a two-state variable that is read and written
 instead of a four-state variable.


-Functions & Tasks
-----------------
-
-All functions and tasks will be inlined (will not become functions in C.)
-The only support provided is simple statements in tasks (which may
-affect global variables).
-
-Recursive functions and tasks are not supported.  All inputs and outputs
-are automatic as if they had the Verilog 2001 "automatic" keyword
-prepended.  (If you don't know what this means, Verilator will do what you
-probably expect, what C does. The default behavior of Verilog is
-different.)
-
-
 Gate Primitives
 ---------------

@@ -383,24 +367,6 @@ Verilator does not support SEREs yet.  All assertion and coverage
 statements must be simple expressions that complete in one cycle.


-Force statement
---------------
-
-Verilator supports the procedural `force` (and corresponding `release`)
-statement. However, the behavior of the `force` statement does not entirely
-comply with IEEE 1800. According to the standard,
-when a procedural statement of the form `force a = b;` is executed, the
-simulation should behave as if, from that point forwards, a continuous
-assignment `assign a = b;` has been added to override the drivers of `a`.
-More specifically: the value of `a` should be updated whenever the value of
-`b` changes, until a `release a;` statement is executed.
-Verilator instead evaluates the current value of `b` when the `force`
-statement is executed, and forces `a` to that value, without updating it
-until a new `force` or `release` statement is encountered that applies to
-`a`. This non-standard behavior is nevertheless consistent with some other
-simulators.
-
-
 Encrypted Verilog
 -----------------

@@ -461,12 +427,31 @@ chandle
  Treated as a "longint"; does not yet warn about operations specified as
  illegal on chandles.

+checker
+  Treated as a "module"; does not yet warn about many constructs illegal
+  inside a checker.
+
 disable
  Disable statements may be used only if the block being disabled is a
  block the disable statement itself is inside.  This was commonly used to
  provide loop break and continue functionality before SystemVerilog added
  the break and continue keywords.

+force, release
+  Verilator supports the procedural `force` (and corresponding `release`)
+  statement. However, the behavior of the `force` statement does not
+  entirely comply with IEEE 1800. According to the standard, when a
+  procedural statement of the form `force a = b;` is executed, the
+  simulation should behave as if, from that point forwards, a continuous
+  assignment `assign a = b;` has been added to override the drivers of `a`.
+  More specifically: the value of `a` should be updated whenever the value
+  of `b` changes, until a `release a;` statement is executed.  Verilator
+  instead evaluates the current value of `b` when the `force` statement is
+  executed, and forces `a` to that value, without updating it until a new
+  `force` or `release` statement is encountered that applies to `a`. This
+  non-standard behavior is nevertheless consistent with some other
+  simulators.
+
 inside
  Inside expressions may not include unpacked array traversal or $ as an
  upper bound.  Case inside and case matches are also unsupported.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 ********
@@ -44,7 +44,7 @@ The best place to get started is to try the :ref:`Examples`.
 .. [#] Verilog is defined by the `Institute of Electrical and Electronics
       Engineers (IEEE) Standard for Verilog Hardware Description
       Language`, Std. 1364, released in 1995, 2001, and 2005.  The
-       Verilator documentation uses the shorthand, e.g., "IEEE 1394-2005",
+       Verilator documentation uses the shorthand, e.g., "IEEE 1364-2005",
       to refer to the, e.g., 2005 version of this standard.

 .. [#] SystemVerilog is defined by the `Institute of Electrical and
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 .. _Simulating:
@@ -143,7 +143,7 @@ Functional Coverage
 -------------------

 With :vlopt:`--coverage` or :vlopt:`--coverage-user`, Verilator will
-translate functional coverage points the user has inserted manually win
+translate functional coverage points the user has inserted manually in
 SystemVerilog code through into the Verilated model.

 Currently, all functional coverage points are specified using SystemVerilog
@@ -498,3 +498,20 @@ Clang.  The process in GCC 10 was as follows:
 Clang and GCC also support -fauto-profile, which uses sample-based
 feedback-directed optimization.  See the appropriate compiler
 documentation.
+
+
+.. _Runtime Debugging:
+
+Runtime Debugging
+=================
+
+To debug a Verilated executable, use the standard GNU debugger ``gdb`` or a
+similar tool. Typically you will want to have debugger symbols inserted by
+the compiler, assertions enabled in the C library, and assertions enabled
+in the Verilated library. (These options slow down the executable, so do
+this only when debugging.) To enable this, Verilate with:
+
+    -CFLAGS -ggdb  -LDFLAGS -ggdb  -CFLAGS -DVL_DEBUG=1  -CFLAGS -D_GLIBCXX_DEBUG
+
+The :vlopt:`-CFLAGS` and/or :vlopt:`-LDFLAGS` options pass arguments
+directly to the compiler or linker.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 **********
@@ -7,9 +7,13 @@ Verilating

 Verilator may be used in five major ways:

+* With the :vlopt:`--binary` option, Verilator will translate the design
+  into an executable, via generating C++ and compiling it.  See
+  :ref:`Binary, C++ and SystemC Generation`.
+
 * With the :vlopt:`--cc` or :vlopt:`--sc` options, Verilator will translate
-  the design into C++ or SystemC code, respectively.  See :ref:`C++ and
-  SystemC Generation`.
+  the design into C++ or SystemC code, respectively.  See :ref:`Binary, C++
+  and SystemC Generation`.

 * With the :vlopt:`--lint-only` option, Verilator will lint the design to
  check for warnings but will not typically create any output files.
@@ -24,13 +28,15 @@ Verilator may be used in five major ways:
  expanded.


-.. _C++ and SystemC Generation:
+.. _Binary, C++ and SystemC Generation:

-C++ and SystemC Generation
-==========================
+Binary, C++ and SystemC Generation
+==================================

 Verilator will translate a SystemVerilog design into C++ with the
-:vlopt:`--cc` option, or into SystemC with the :vlopt:`--sc` option.
+:vlopt:`--cc` option, or into SystemC with the :vlopt:`--sc` option.  It
+will translate into C++ and compile it into an executable binary with the
+:vlopt:`--binary` option.

 When using these options:

@@ -45,14 +51,19 @@ When using these options:
   The prefix is set with :vlopt:`--prefix`, or defaults to the name of the
   top module.

-#. If :vlopt:`--exe` is used, Verilator creates makefiles to generate a
-   simulation executable, otherwise, it creates makefiles to generate an
-   archive (.a) containing the objects.
+#. If :vlopt:`--binary` or :vlopt:`--main` is used, Verilator creates a C++
+   top wrapper to read command line arguments, create the model, and
+   execute the model.

-#. If :vlopt:`--build` option was used, it calls :ref:`GNU Make` or
-   :ref:`CMake` to build the model.
+#. If :vlopt:`--binary` or :vlopt:`--exe` is used, Verilator creates
+   makefiles to generate a simulation executable, otherwise, it creates
+   makefiles to generate an archive (.a) containing the objects.

-Once a model is built, it is then typically run, see :ref:`Simulating`.
+#. If :vlopt:`--binary` or :vlopt:`--build` is used, it calls :ref:`GNU
+   Make` or :ref:`CMake` to build the model.
+
+Once a model is built, the next step is typically for the user to run it,
+see :ref:`Simulating`.


 .. _Hierarchical Verilation:
@@ -165,7 +176,7 @@ compiled on the different target system.

 To support this, none of the files that Verilator produces will reference
 any configure-generated build-system-specific files, such as
-:file:`config.h` (which is renamed in Verilator to :file:`config_build.h`
+:file:`config.h` (which is renamed in Verilator to :file:`config_package.h`
 to reduce confusion.)  The disadvantage of this approach is that
 :file:`include/verilatedos.h` must self-detect the requirements of the
 target system, rather than using configure.
@@ -1,4 +1,4 @@
-.. Copyright 2003-2023 by Wilson Snyder.
+.. Copyright 2003-2024 by Wilson Snyder.
 .. SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

 *******************
@@ -12,10 +12,18 @@ Disabling Warnings

 Warnings may be disabled in multiple ways:

-#. Disable the warning in the source code.  When the warning is printed, it
-   will include a warning code.  Surround the offending line with a
-   :code:`/*verilator&32;lint_off*/` and :code:`/*verilator&32;lint_on*/`
-   metacomment pair:
+#. Disable the warning globally by invoking Verilator with the
+   :code:`-Wno-{warning-code}` option.
+
+   Global disables should be avoided, as they removes all checking across
+   the source files, and prevents other users from compiling the sources
+   without knowing the magic set of disables needed to compile those
+   sources successfully.
+
+#. Disable the warning in the design source code.  When the warning is
+   printed, it will include a warning code.  Surround the offending line
+   with a :code:`/*verilator&32;lint_off*/` and
+   :code:`/*verilator&32;lint_on*/` metacomment pair:

   .. code-block:: sv

@@ -23,6 +31,12 @@ Warnings may be disabled in multiple ways:
         if (`DEF_THAT_IS_EQ_ZERO <= 3) $stop;
         // verilator lint_on UNSIGNED

+
+   A lint_off in the design source code will propagate down to any child
+   files (files later included by the file with the lint_off), but will not
+   propagate upwards to any parent file (file that included the file with
+   the lint_off).
+
 #. Disable the warning using :ref:`Configuration Files` with a
   :option:`lint_off` command.  This is useful when a script suppresses
   warnings, and the Verilog source should not be changed.  This method also
@@ -32,12 +46,6 @@ Warnings may be disabled in multiple ways:

         lint_off -rule UNSIGNED -file "*/example.v" -line 1

-#. Disable the warning globally invoking Verilator with the
-   :code:`-Wno-{warning-code}` option.  This should be avoided, as it
-   removes all checking across the designs, and prevents other users from
-   compiling your code without knowing the magic set of disables needed to
-   compile your design successfully.
-

 Error And Warning Format
 ========================
@@ -74,7 +82,8 @@ List Of Warnings
 .. option:: Unsupported: ....

   This error indicates that the code uses a Verilog language construct
-   that is not yet supported in Verilator.  See the Limitations chapter.
+   that is not yet supported in Verilator.  See also :ref:`Language
+   Limitations`.


 .. option:: ALWCOMBORDER
@@ -96,6 +105,24 @@ List Of Warnings
   simulate correctly.


+.. option:: ASCRANGE
+
+   .. TODO better example
+
+   Warns that a packed vector is declared with ascending bit range
+   (i.e. [0:7]).  Descending bit range is now the overwhelming standard,
+   and ascending ranges are now thus often due to simple oversight
+   instead of intent (a notable exception is the OpenPOWER code base).
+
+   It also warns that an instance is declared with ascending range
+   (i.e. [0:7] or [7]) and is connected to an N-wide signal.
+   The bits will likely be in the reversed order from what people may expect
+   (i.e., instance [0] will connect to signal bit [N-1] not bit [0]).
+
+   Ignoring this warning will only suppress the lint check; it will
+   simulate correctly.
+
+
 .. option:: ASSIGNDLY

   .. TODO better example
@@ -317,9 +344,11 @@ List Of Warnings

 .. option:: CDCRSTLOGIC

-   With :vlopt:`--cdc` only, it warns that asynchronous flop reset terms come
-   from other than primary inputs or flopped outputs, creating the
-   potential for reset glitches.
+   Historical, never issued since version 5.008.
+
+   Warned with a no longer supported clock domain crossing option that
+   asynchronous flop reset terms came from other than primary inputs or
+   flopped outputs, creating the potential for reset glitches.


 .. option:: CLKDATA
@@ -375,6 +404,16 @@ List Of Warnings
   simulators.


+.. option:: CONSTRAINTIGN
+
+   Warns that Verilator does not support :code:`constraint`,
+   :code:`constraint_mode`, or :code:`rand_mode`, and the construct was are
+   ignored.
+
+   Ignoring this warning may make Verilator randomize() simulations differ
+   from other simulators.
+
+
 .. option:: CONTASSREG

   .. TODO better example
@@ -639,6 +678,52 @@ List Of Warnings
   used as a clock.


+.. option:: GENUNNAMED
+
+   Warns that a generate block was unnamed and "genblk" will be used per
+   IEEE.
+
+   The potential issue is that adding additional generate blocks will
+   renumber the assigned names, which may cause eventual problems with
+   synthesis constraints or other tools that depend on hierarchical paths
+   remaining consistent.
+
+   Blocks that are empty may not be reported with this warning, as no
+   scopes are created for empty blocks, so there is no harm in having them
+   unnamed.
+
+   Disabled by default as this is a code-style warning; it will simulate
+   correctly.
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 2
+
+         generate
+            if (PARAM == 1) begin  //<--- Warning
+            end
+
+   Results in:
+
+   .. code-block::
+
+         %Warning-GENUNNAMED: example.v:2:9: Unnamed generate block (IEEE 1800-2017 27.6)
+
+   To fix this assign a label (often with the naming convention prefix of
+   :code:`gen_` or :code:`g_`), for example:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 2
+
+         generate
+            if (PARAM == 1) begin : gen_param_1  //<--- Repaired
+            end
+
+   Other tools with similar warnings: Verible's generate-label, "All
+   generate block statements must have a label."
+
+
 .. option:: HIERBLOCK

   Warns that the top module is marked as a hierarchy block by the
@@ -735,14 +820,17 @@ List Of Warnings

   This is a warning because the static default differs from C++, differs
   from class member function/tasks.  Static is a more dangerous default
-   then automatic as static prevents the function from being reinterant,
+   then automatic as static prevents the function from being reentrant,
   which may be a source of bugs, and/or performance issues.

-   If the function does not require static behavior, change it to "function
-   automatic".
+   If the function is in a module, and does not require static behavior,
+   change it to "function automatic".

-   If the function requires static behavior, change it to "function
-   static".
+   If the function is in a module, and requires static behavior, change it
+   to "function static".
+
+   If the function is in a package, it defaults to static, and label the
+   function's variables as static.

   Ignoring this warning will only suppress the lint check; it will
   simulate correctly.
@@ -842,23 +930,111 @@ List Of Warnings
   Ignoring this warning will only suppress the lint check; it will
   simulate correctly.

+.. option:: LIFETIME
+
+   Error when a variable is referenced in a process that can outlive the process
+   in which it was declared. This can happen when using 'fork..join_none' or
+   'fork..join_any' blocks, which spawn process that can outlive their parents.
+   This error occurs only when Verilator can't replace the reference with a
+   reference to copy of this variable, local to the forked process. For example:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 3
+
+         task foo(int local_var);
+            fork
+               #10 local_var++;
+               #20 $display("local_var = %d", local_var);
+            join_none
+         endtask
+
+   In the example above 'local_var' exists only within scope of 'foo', once foo
+   finishes, the stack frame containing 'i' gets removed. However, the process
+   forked from foo continues, as it contains a delay. After 10 units of time
+   pass, this process attempts to modify 'local_var'. However, this variable no
+   longer exits. It can't be made local to the forked process upon spawning, because
+   it's modified and can be referenced somewhere else, for example in the other
+   forked process, that was delayed by 20 units of time in this example. Thus,
+   there's no viable stack allocation for it.
+
+   In order to fix it, if the intent is not to share the variable's state outside
+   of the process, then create a local copy of the variable.
+
+   For example:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 4
+
+         task foo(int local_var);
+            fork
+               #10 begin
+                  int forked_var = local_var;
+                  forked_var++;
+               end
+               #20 begin
+                  // Note that we are going to print the original value here,
+                  // as `forked_var`is a local copy that was initialized while
+                  // `foo` was still alive.
+                  int forked_var = local_var;
+                  $display("forked_var = %d", forked_var)
+               end
+            join_none
+         endtask
+
+   If you need to share its state, another strategy is to ensure it's allocated
+   statically:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 1
+
+         int static_var;
+
+         task foo();
+            fork
+               #10 static_var++;
+               #20 $display("static_var = %d", static_var);
+            join_none
+         endtask
+
+   However, if you need to be able to instantiate at runtime, the solution would be to
+   wrap it in an object, since the forked process can hold a reference to that object
+   and ensure that the variable stays alive this way:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 2
+
+         class Wrapper;
+            int m_var;
+
+            // Here we implicitly hold a reference to `this`
+            task foo();
+               fork
+                  #10 m_var++;
+                  #20 $display("this.m_var = %d", m_var);
+               join_none
+            endtask
+         endclass
+
+         // Here we explicitly hold a handle to an object
+         task bar(Wrapper wrapper);
+            fork
+               #10 wrapper.m_var++;
+               #20 $display("wrapper.m_var = %d", wrapper.m_var);
+            join_none
+         endtask

 .. option:: LITENDIAN

   .. TODO better example

-   Warns that a packed vector is declared with little endian bit numbering
-   (i.e. [0:7]).  Big endian bit numbering is now the overwhelming
-   standard, and little numbering is now thus often due to simple oversight
-   instead of intent.
-
-   It also warns that an instance is declared with little endian range
-   (i.e. [0:7] or [7]) and is connected to an N-wide signal. Based on IEEE
-   the bits will likely be backward from what people may expect
-   (i.e., instance [0] will connect to signal bit [N-1] not bit [0]).
-
-   Ignoring this warning will only suppress the lint check; it will
-   simulate correctly.
+   The naming of this warning is in contradiction with the common
+   interpretation of little endian. It was therefore renamed to
+   :option:`ASCRANGE`. While :option:`LITENDIAN` remains for
+   backwards compatibility, new projects should use :option:`ASCRANGE`.


 .. option:: MINTYPMAX
@@ -871,6 +1047,51 @@ List Of Warnings
   unsupported. Verilator uses only the typical delay value.


+.. option:: MISINDENT
+
+   Warns that the indentation of a statement is misleading, suggesting the
+   statement is part of a previous :code:`if` or :code:`while` block while
+   it is not.
+
+   Verilator suppresses this check when there is an inconsistent mix of
+   spaces and tabs, as it cannot ensure the width of tabs.  Verilator also
+   ignores blocks with :code:`begin`/:code:`end`, as the :code:`end`
+   visually indicates the earlier statement's end.
+
+   Ignoring this warning will only suppress the lint check; it will
+   simulate correctly.
+
+   For example
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 3
+
+         if (something)
+            statement_in_if;
+            statement_not_in_if;  //<--- Warning
+
+   Results in:
+
+   .. code-block::
+
+         %Warning-MISINDENT: example.v:3:9: Misleading indentation
+
+   To fix this repair the indentation to match the correct earlier
+   statement, for example:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 3
+
+         if (something)
+            statement_in_if;
+         statement_not_in_if;  //<--- Repaired
+
+   Other tools with similar warnings: GCC -Wmisleading-indentation,
+   clang-tidy readability-misleading-indentation.
+
+
 .. option:: MODDUP

   .. TODO better example
@@ -949,6 +1170,21 @@ List Of Warnings
   (neither :vlopt:`--timing` nor :vlopt:`--no-timing` option was provided).


+.. option:: NEWERSTD
+
+   Warns that a feature requires a newer standard of Verilog or SystemVerilog
+   than the one specified by the :vlopt:`--language` option. For example, unsized
+   unbased literals (`'0`, `'1`, `'z`, `'x`) require IEEE 1800-2005 or later.
+
+   To avoid this warning, use a Verilog or SystemVerilog standard that
+   supports the feature. Alternatively, modify your code to use a different
+   syntax that is supported by the Verilog/SystemVerilog standard specified
+   by the :vlopt:`--language` option.
+
+   Ignoring this warning will only suppress the lint check; it will
+   simulate correctly.
+
+
 .. option:: NOLATCH

   .. TODO better example
@@ -1132,7 +1368,10 @@ List Of Warnings

 .. option:: RANDC

-   Warns that the :code:`randc` keyword is unsupported and being converted
+   Historical, never issued since version 5.018, when :code:`randc` became
+   fully supported.
+
+   Warned that the :code:`randc` keyword was unsupported and was converted
   to :code:`rand`.


@@ -1275,6 +1514,39 @@ List Of Warnings
   modeled values, or DPI calls.


+.. option:: SIDEEFFECT
+
+   Warns that an expression has a side effect that might not properly be
+   executed by Verilator.
+
+   This often represents a bug in Verilator, as opposed to a bad code
+   construct, however the Verilog code can typically be changed to avoid
+   the warning.
+
+   Faulty example:
+
+   .. code-block:: sv
+      :linenos:
+      :emphasize-lines: 1
+
+         x = y[a++];
+
+   This example warns because Verilator does not currently handle side
+   effects inside array subscripts; the a++ may be executed multiple times.
+
+   Rewrite the code to avoid expression side effects, typically by using a
+   temporary:
+
+   .. code-block:: sv
+      :linenos:
+
+         temp = a++;
+         x = y[temp];
+
+   Ignoring this warning may make Verilator simulations differ from other
+   simulators.
+
+
 .. option:: SPLITVAR

   Warns that a variable with a :option:`/*verilator&32;split_var*/`
@@ -1299,6 +1571,17 @@ List Of Warnings
   * The variable is tristate or bidirectional. (e.g., :code:`inout`).


+.. option:: STATICVAR
+
+   Warns that a static variable declared in a loop with declaration assignment
+   was converted to automatic. Often such variables were intended to
+   instead be declared "automatic".
+
+   Ignoring this warning may make Verilator differ from other simulators,
+   which will treat the variable as static. Verilator may in future versions also
+   treat the variable as static.
+
+
 .. option:: STMTDLY

   Warns that the code has a statement with a delayed time in front of it.
@@ -1751,11 +2034,15 @@ List Of Warnings

   .. code-block:: sv

-         wait(0);  // Blocks forever
+         wait(1);  // Blocks forever

   Warns that a `wait` statement awaits a constant condition, which means it
   either blocks forever or never blocks.

+   As a special case `wait(0)` with the literal constant `0` (as opposed to
+   something that elaborates to zero), does not warn, as it is presumed the
+   code is making the intent clear.
+

 .. option:: WIDTH

@@ -1790,17 +2077,29 @@ List Of Warnings

   For example, this is a missized index:

-   .. include:: ../../docs/gen/ex_WIDTH_1_faulty.rst
+   .. include:: ../../docs/gen/ex_WIDTHEXPAND_1_faulty.rst

-   Results in:
+   Results in a WIDTHEXPAND warning:

-   .. include:: ../../docs/gen/ex_WIDTH_1_msg.rst
+   .. include:: ../../docs/gen/ex_WIDTHEXPAND_1_msg.rst

   One possible fix:

-   .. include:: ../../docs/gen/ex_WIDTH_1_fixed.rst
+   .. include:: ../../docs/gen/ex_WIDTHEXPAND_1_fixed.rst


+.. option:: WIDTHTRUNC
+
+   A more granular WIDTH warning, for when a value is truncated.
+
+.. option:: WIDTHEXPAND
+
+   A more granular WIDTH warning, for when a value is zero expanded.
+
+.. option:: WIDTHXZEXPAND
+
+   A more granular WIDTH warning, for when a value is X/Z expanded.
+
 .. option:: WIDTHCONCAT

   Warns that based on the width rules of Verilog, a concatenate, or
@@ -628,21 +628,32 @@ coroutines ``co_await`` its ``join`` function, and forked ones call ``done``
 when they're finished. Once the required number of coroutines (set using
 ``setCounter``) finish execution, the forking coroutine is resumed.

-Awaitable Utilities
-~~~~~~~~~~~~~~~~~~~
+``VlForever``
+~~~~~~~~~~~~~

-There are also two small utility awaitable types:
-
-* ``VlNow`` is an awaitable that suspends and immediately resumes coroutines.
-  It is used for forcing a coroutine to be moved onto the heap. See the `Forks`
-  section for more detail.
-* ``VlForever`` is used for blocking a coroutine forever. See the `Timing pass`
-  section for more detail.
+A small utility awaitable type. It allows for blocking a coroutine forever. It
+is currently only used for ``wait`` statements that await a constant false
+condition. See the `Timing Pass` section for more details.

 Timing Pass
 ~~~~~~~~~~~

-The visitor in ``V3Timing.cpp`` transforms each timing control into a ``co_await``.
+There are two visitors in ``V3Timing.cpp``.
+
+The first one, ``TimingSuspendableVisitor``, does not perform any AST
+transformations. It is responsible for marking processes and C++ functions that
+contain timing controls as suspendable. Processes that call suspendable
+functions are also marked as suspendable. Functions that call, are overridden
+by, or override suspendable functions are marked as suspendable as well.
+
+The visitor keeps a dependency graph of functions and processes to handle such
+cases. A function or process is dependent on a function if it calls it. A
+virtual class method is dependent on another class method if it calls it,
+overrides it, or is overriden by it.
+
+The second visitor in ``V3Timing.cpp``, ``TimingControlVisitor``, uses the
+information provided by ``TimingSuspendableVisitor`` and transforms each timing
+control into a ``co_await``.

 * event controls are turned into ``co_await`` on a trigger scheduler's
  ``trigger`` method. The awaited trigger scheduler is the one corresponding to
@@ -674,14 +685,9 @@ Each sub-statement of a ``fork`` is put in an ``AstBegin`` node for easier
 grouping. In a later step, each of these gets transformed into a new, separate
 function. See the `Forks` section for more detail.

-Processes that use awaits are marked as suspendable. Later, during ``V3Sched``,
-they are transformed into coroutines. Functions that use awaits get the return
-type of ``VlCoroutine``. This immediately makes them coroutines. Note that if a
-process calls a function that is a coroutine, the call gets wrapped in an
-await, which means the process itself will be marked as suspendable. A virtual
-function is a coroutine if any of its overriding or overridden functions are
-coroutines. The visitor keeps a dependency graph of functions and processes to
-handle such cases.
+Suspendable functions get the return type of ``VlCoroutine``, which makes them
+coroutines. Later, during ``V3Sched``, suspendable processes are also
+transformed into coroutines.

 Scheduling with timing
 ~~~~~~~~~~~~~~~~~~~~~~
@@ -749,9 +755,7 @@ doesn't suspend the forking process.

 In forked processes, references to local variables are only allowed in
 ``fork..join``, as this is the only case that ensures the lifetime of these
-locals are at least as long as the execution of the forked processes. This is
-where ``VlNow`` is used, to ensure the locals are moved to the heap before they
-are passed by reference to the forked processes.
+locals are at least as long as the execution of the forked processes.


 Multithreaded Mode
@@ -1031,11 +1035,11 @@ Coding Conventions
 ==================


-Compiler Version and C++11
+Compiler Version and C++14
 --------------------------

-Verilator requires C11. Verilator does not require any newer versions, but
-is maintained to build successfully with C14/C17/C20.
+Verilator requires C14. Verilator does not require any newer versions, but
+is maintained to build successfully with C17/C20.


 Indentation and Naming Style
@@ -1218,9 +1222,9 @@ There are three ways data is passed between visitor functions.

 2. User attributes. Each ``AstNode`` (**Note.** The AST node, not the
   visitor) has five user attributes, which may be accessed as an
-   integer using the ``user1()`` through ``user5()`` methods, or as a
+   integer using the ``user1()`` through ``user4()`` methods, or as a
   pointer (of type ``AstNUser``) using the ``user1p()`` through
-   ``user5p()`` methods (a common technique lifted from graph traversal
+   ``user4p()`` methods (a common technique lifted from graph traversal
   packages).

   A visitor first clears the one it wants to use by calling
@@ -1396,7 +1400,7 @@ Controlling the Test Driver
 ---------------------------

 The test driver script `driver.pl` runs tests; see the `Test Driver`
-section.  The individual test drivers are written in PERL; see `Test
+section.  The individual test drivers are written in Perl; see `Test
 Language`.


@@ -1428,7 +1432,7 @@ Developers will also want to call ./configure with two extra flags:
   disabled by default, as SystemC installation problems would otherwise
   falsely indicate a Verilator problem.

-When enabling the long tests, some additional PERL modules are needed,
+When enabling the long tests, some additional Perl modules are needed,
 which you can install using cpan.

 ::
@@ -1458,6 +1462,19 @@ There are some traps to avoid when running regression tests
  system-dependent, but on Fedora Linux it would require editing the
  ``/etc/security/limits.conf`` file as root.

+Diffing generated code after changes
+------------------------------------
+
+When making a change in the code generation area that should not change the
+actual emitted code, it is useful to perform a diff to make sure the emitted
+code really did not change. To do this, the top level Makefile provides the
+*test-snap* and *test-diff* targets:
+
+- Run the test suite with ``make test``
+- Take a snapshot with ``make test-snap``
+- Apply your changes
+- Run the test suite again with ``make test``
+- See the changes in the output with ``make test-diff``

 Continuous Integration
 ----------------------
@@ -1509,8 +1526,19 @@ debug level 5, with the V3Width.cpp file at level 9.
 --debug
 -------

-When you run with ``--debug``, there are two primary output file types
-placed into the obj_dir, .tree and .dot files.
+When you run with ``--debug``, there are three primary output file types
+placed into the obj_dir, .vpp, .tree and .dot files.
+
+.vpp Output
+-----------
+
+Verilator creates a *{mod_prefix}*\ __inputs\ .vpp file containing all the
+files that were read, filtered by preprocessing. This file can be fed back
+into Verilator, replacing on the command line all of the previous input
+files, to enable simplification of test cases.
+
+Verilator also creates .vpp files for each individual file passed on the
+command line.


 .dot Output
@@ -1625,7 +1653,7 @@ Source file and line
   file is ``a``, the 26th is ``z``, the 27th is ``aa``, and so on.

 User pointers
-   Shows the value of the node's user1p...user5p, if non-NULL.
+   Shows the value of the node's user1p...user4p, if non-NULL.

 Data type
   Many nodes have an explicit data type. "@dt=0x..." indicates the
@@ -1807,8 +1835,6 @@ IEEE 1800-2017 6.12 "shortreal"
    Little/no tool support, and easily promoted to real.
 IEEE 1800-2017 11.11 Min, typ, max
    No SDF support, so will always use typical.
-IEEE 1800-2017 11.12 "let"
-    Little/no tool support, makes it difficult to implement parsers.
 IEEE 1800-2017 20.16 Stochastic analysis
    Little industry use.
 IEEE 1800-2017 20.17 PLA modeling
@@ -2110,7 +2136,7 @@ Or in a hand-written C++ wrapper:
 ::

   #ifdef TEST_VERBOSE
-      cout << "Read a=" << a << endl;
+      std::cout << "Read a=" << a << std::endl;
   #endif

 A filename that should be used to check the output results is given with
@@ -2211,7 +2237,7 @@ xsim_flags / xsim_flags2 / xsim_run_flags
 Distribution
 ============

-Copyright 2008-2023 by Wilson Snyder. Verilator is free software; you can
+Copyright 2008-2024 by Wilson Snyder. Verilator is free software; you can
 redistribute it and/or modify it under the terms of either the GNU Lesser
 General Public License Version 3 or the Perl Artistic License Version 2.0.

@@ -1,4 +1,5 @@
 ABCp
+Aadi
 Accellera
 Affe
 Aleksander
@@ -6,27 +7,33 @@ Alexandre
 Ami
 Amir
 Anastasiadis
+Anglin
 Anikin
 Antmicro
 Antonin
 Antwerpen
 Arasanipalai
 Arjen
+Arshid
 Asciidoc
 Ashutosh
 Ast
 Atmel
 Aurelien
 Bagri
+Balboni
 Baltazar
 Bardsley
 Benchmarking
 Berman
 Bhagavatula
 Bhattacharyya
+Biancolin
 Binderman
+Björk
 Bleiweiss
 Bogdan
+Bonneau
 Bouchard
 Boumenot
 Brej
@@ -34,29 +41,40 @@ Briquet
 Brownlee
 Buckenmaier
 Bybell
+CLang
 CMake
 Cai
 Cavium
 Cfuncs
 Chandan
+Chentai
+Cherkasov
+Chih
 Chitlesh
 Christophe
 Cochrane
+Conda
+Corteggiani
+Cozzocrea
 Cuan
 Cygwin
 DErrico
 DFG
 Da
+Dai
 Danilo
 Dannoritzer
 Das
+Davide
 Dcache
 Deadman
 Debacker
+Deepa
 Defparams
 Denio
 Deprecations
 Deroo
+Desai
 Dickol
 Diez
 Dimitris
@@ -65,9 +83,11 @@ Djordjevic
 Dobbie
 Dockerfile
 Donal
+Donovick
 Doorn
 Doxygen
 Dressler
+Dudek
 Duraid
 Dutton
 Dzetkulic
@@ -85,17 +105,24 @@ Engstrom
 Enzo
 Eugen
 Fabrizio
+Faucher
+Faure
 Fekete
 Ferrandi
 Flachs
 Flavien
+Florian
 Foletto
+Forencich
 Forker
 Francillon
 Fredieu
+Freiberger
+Frédéric
 GTKWave
 Galbi
 Gantt
+Garnal
 Gartner
 Gelinek
 Geoff
@@ -111,18 +138,23 @@ Goessling
 Gonnen
 Goorah
 Gossner
+Graba
 Graphviz
 Graybeal
 Grobman
 Grulfen
+Gu
 Gunter
 Guo
 Hao
+Haojin
 Harboe
 Hashimoto
 Hayenga
+Hesselbarth
 Hewson
 Hicamp
+Hideto
 Hiroki
 Holdsworth
 Holme
@@ -130,8 +162,10 @@ Homebrew
 Hornung
 Hossell
 Hsu
+Hutt
 Hyperthreading
 Ibrahim
+Ikram
 Iles
 Inlines
 Inout
@@ -143,6 +177,7 @@ Jae
 Jalloq
 Jankowski
 Jannis
+Jasen
 Jens
 Jeras
 Jiuyang
@@ -158,6 +193,7 @@ Kahn
 Kai
 Kamendje
 Kandadi
+Kaneda
 Kaplan
 Karge
 Karlsson
@@ -173,6 +209,7 @@ Klnv
 Koenig
 Koh
 Kolecki
+Konstantin
 Koonce
 Korteland
 Koszek
@@ -184,16 +221,23 @@ Kumashikar
 Kuoping
 Kurachi
 Laeufer
+Laroche
 Laurens
 Lavino
 Leber
 Leendert
+Lem
 Lesik
+Libenzi
+Licker
 Liland
 Liu
+Liwei
 Lockhart
 Longo
+Luca
 Lussier
+Lübeck
 MMD
 MODDUP
 MTasks
@@ -203,10 +247,17 @@ Mahesh
 Mahmoudy
 Makefile
 Makefiles
+Maksymenko
+Markley
+Marquet
 Matveyenko
 Maupin
 Mdir
 Mednick
+Mei
+Melo
+Menküc
+Michail
 Michiels
 Microsystems
 Milanovic
@@ -222,9 +273,13 @@ Mong
 Muhlestein
 Multithreaded
 Multithreading
+Mykyta
 NOUNOPTFLAT
+NaN
 Nalbantis
+Nandor
 Narayan
+Nassim
 Nauticus
 Newgard
 Nigam
@@ -233,15 +288,20 @@ Niranjan
 Nitza
 Noack
 Nodine
+Ober
 Oleg
 Olof
 Olofsson
+Oron
 Oyvind
 PLI
+Palaniappan
 Patricio
 Petr
+Piard
 Piechotka
 Piersall
+Platz
 Platzer
 Plunkett
 Popolon
@@ -252,10 +312,14 @@ Prateek
 Pre
 Preprocess
 Pretet
+Pretl
+Pringlemeir
 Priyadharshini
 Pullup
 Pulver
 Puri
+Qin
+Qiu
 Questa
 Rachit
 Ralf
@@ -265,27 +329,37 @@ Reitan
 Renga
 Requin
 Rodionov
+Rohan
 Rolfe
+Rontionov
 Roodselaar
+Roshit
 Runtime
 Ruud
 Rystsov
 STandarD
 Salman
+Sammelson
+Sanggyu
 Sanguinetti
 Sanjay
+Sasselli
 Scharrer
 Seitz
+Shahid
 Shankar
 Shanshan
 Sharad
 Shareef
+Shen
 Sheng
 Shi
 Shinkarovsky
+Shinya
 Shirakawa
 Shuba
 Shunyao
+Slager
 Slatter
 SoC
 Sobhan
@@ -295,6 +369,7 @@ Solomatnikov
 Solt
 Southwell
 Srini
+Srinivasan
 Stamness
 Stephane
 Stoddard
@@ -304,6 +379,7 @@ Stucki
 Su
 Suguimoto
 Sundararajan
+Sunderland
 Suse
 Syms
 Synopsys
@@ -311,6 +387,7 @@ SystemC
 SystemVerilog
 Takatsukasa
 Tambe
+Tarik
 Tariq
 Tejada
 Tengstrand
@@ -318,7 +395,9 @@ Terpstra
 Thiede
 Thierry
 Thyer
+Tianrui
 Tichelaar
+Timi
 Tomov
 Tood
 Topa
@@ -327,9 +406,11 @@ Trefor
 Tresidder
 Tri
 Tristate
+Truong
 Turton
 UNOPTFLAT
 Ubixum
+Ueno
 Uints
 Undefines
 Unsized
@@ -338,11 +419,13 @@ Uselib
 Usha
 Usuario
 VERILATOR
+Vandergriendt
 Vasu
 Vdeeptemp
 Vdly
 Vemumtab
 Vemuri
+Venkataramanan
 Veriable
 Verialted
 Verilate
@@ -353,6 +436,7 @@ Verilator
 Verilog
 Vighnesh
 Viktor
+Vilp
 Vm
 Vukobratovic
 Wai
@@ -363,6 +447,7 @@ Wfuture
 Whatson
 Wildman
 Wim
+Wmisleading
 Wno
 Wojciech
 Wolfel
@@ -372,13 +457,20 @@ Wwarn
 XSim
 Xcelium
 Xiaoliang
+Xiaoyi
 Xuan
 Xuanqi
 Yazdanbakhsh
+Yernagula
 Yi
+Yike
 Yinan
+Yoshitomo
 Yosys
 Yu
+Yujia
+Yurii
+Zaruba
 Zhang
 abirkmanis
 accessor
@@ -396,7 +488,9 @@ architected
 args
 arrarys
 assertOn
+astgen
 async
+atClone
 ato
 atoi
 autoconf
@@ -452,16 +546,19 @@ commandArgsPlusMatch
 compilable
 concat
 concats
+conf
 config
 const
 constexpr
 constpool
 coredump
 coroutine
+coroutines
 countbits
 countones
 cout
 cpp
+cppstyle
 cpu
 ctor
 ctrace
@@ -488,6 +585,7 @@ desassign
 destructor
 detections
 dev
+devcontainer
 devel
 dir
 displayb
@@ -496,6 +594,7 @@ doxygen
 dpiGetContext
 dpic
 dsvf
+dt
 dtor
 dumpall
 dumpfile
@@ -512,6 +611,7 @@ dumpportson
 dumpvars
 dut
 dx
+dynarray
 elab
 elike
 elsif
@@ -520,6 +620,7 @@ endcelldefine
 endfunction
 endgenerate
 endian
+endianness
 endif
 endmodule
 endprotect
@@ -541,6 +642,7 @@ exe
 executables
 expr
 extern
+ezchi
 fanin
 fasttrace
 fauto
@@ -585,6 +687,7 @@ genvar
 genvars
 getenv
 getline
+ggdb
 gmake
 gmon
 gotFinish
@@ -629,16 +732,19 @@ ish
 isunbounded
 isunknown
 jobserver
+jwoutersymatra
 killua
 lang
 lcov
 ld
 leavinel
 len
+libc
 libext
 libgoogle
 libsystemc
 libtcmalloc
+libverilated
 linkers
 linter
 linux
@@ -653,7 +759,9 @@ logicals
 longint
 lossy
 lsb
+lubc
 lvalue
+lvalues
 lxt
 macromodule
 makefile
@@ -665,6 +773,7 @@ miree
 mis
 misconnected
 misconversion
+misoptimized
 missized
 mk
 mno
@@ -678,14 +787,18 @@ mtasks
 mulithreaded
 mult
 multidim
+multidriven
 multiinterfaces
 multiline
 multipling
+multipoint
 multithread
 multithreaded
 multithreading
+musl
 mutexes
 mux
+muzafferkal
 myftptoyman
 mysignal
 namespace
@@ -714,6 +827,7 @@ ortegon
 oversubscription
 parallelized
 param
+parameterization
 parameterized
 params
 parens
@@ -723,6 +837,7 @@ pdf
 perf
 perftools
 pgo
+phelter
 picoChip
 pinIndex
 pinout
@@ -740,7 +855,9 @@ pragma
 pragmas
 pre
 precisions
+precompiled
 predefines
+prepareClone
 prepend
 prepended
 preprocess
@@ -762,10 +879,14 @@ pulldowns
 pullup
 pvalue
 pwd
+py
 qrq
+qrqiuren
 radix
 randc
 randcase
+randstate
+raphmaster
 rarr
 rdtsc
 reStructuredText
@@ -775,6 +896,7 @@ readmemb
 readmemh
 realpath
 realtime
+realtobits
 recoding
 recrem
 redeclaring
@@ -793,6 +915,7 @@ sVerilator
 sawatzke
 sc
 scalared
+sccache
 sccanf
 seg
 setuphold
@@ -807,6 +930,7 @@ specparam
 splitme
 spp
 sqrt
+srandom
 src
 srcdir
 srcfile
@@ -838,6 +962,7 @@ svdpi
 swrite
 sys
 systemc
+tenghtt
 testbench
 threadsafe
 threashold
@@ -869,10 +994,12 @@ typedefs
 typename
 uint
 un
+unbased
 undef
 undefineall
 undriven
 ungetc
+unhandled
 uniquified
 unistd
 unlink
@@ -893,7 +1020,9 @@ vc
 vcd
 vcddiff
 vcoverage
+vdhotre
 vec
+ventana
 ver
 verFiles
 verible
@@ -909,6 +1038,9 @@ vlt
 vltstd
 vluint
 vpi
+vpiConstType
+vpiDefName
+vpiInstance
 vpiLeftRange
 vpiModule
 vpiSize
@@ -938,4 +1070,5 @@ ypq
 yurivict
 zdave
 Øyvind
-
+Алексеевич
+Исаак
@@ -70,7 +70,7 @@ The XML document consists of 4 sections within the top level
 Distribution
 ============

-Copyright 2020-2023 by Wilson Snyder. Verilator is free software; you can
+Copyright 2020-2024 by Wilson Snyder. Verilator is free software; you can
 redistribute it and/or modify it under the terms of either the GNU Lesser
 General Public License Version 3 or the Perl Artistic License Version 2.0.

@@ -19,7 +19,8 @@
 # cmake ..
 # cmake --build .

-cmake_minimum_required(VERSION 3.8)
+cmake_minimum_required(VERSION 3.12)
+cmake_policy(SET CMP0074 NEW)
 project(cmake_hello_c)

 find_package(verilator HINTS $ENV{VERILATOR_ROOT} ${VERILATOR_ROOT})
@@ -19,7 +19,8 @@
 # cmake ..
 # cmake --build .

-cmake_minimum_required(VERSION 3.8)
+cmake_minimum_required(VERSION 3.12)
+cmake_policy(SET CMP0074 NEW)
 project(cmake_hello_sc CXX)

 find_package(verilator HINTS $ENV{VERILATOR_ROOT} ${VERILATOR_ROOT})
@@ -19,7 +19,8 @@
 # cmake ..
 # cmake --build .

-cmake_minimum_required(VERSION 3.8)
+cmake_minimum_required(VERSION 3.12)
+cmake_policy(SET CMP0074 NEW)
 project(cmake_protect_lib)

 find_package(verilator HINTS $ENV{VERILATOR_ROOT} ${VERILATOR_ROOT})
@@ -19,7 +19,8 @@
 # cmake ..
 # cmake --build .

-cmake_minimum_required(VERSION 3.8)
+cmake_minimum_required(VERSION 3.12)
+cmake_policy(SET CMP0074 NEW)
 project(cmake_tracing_c)

 find_package(verilator HINTS $ENV{VERILATOR_ROOT} ${VERILATOR_ROOT})
@@ -19,7 +19,9 @@
 # cmake ..
 # cmake --build .

-cmake_minimum_required(VERSION 3.8)
+cmake_minimum_required(VERSION 3.12)
+cmake_policy(SET CMP0074 NEW)
+
 project(cmake_tracing_sc_example CXX)

 find_package(verilator HINTS $ENV{VERILATOR_ROOT} ${VERILATOR_ROOT})
@@ -7,7 +7,7 @@
 //======================================================================

 // SystemC global header
-#include <systemc.h>
+#include <systemc>

 // Include common routines
 #include <verilated.h>
@@ -15,6 +15,8 @@
 // Include model header, generated from Verilating "top.v"
 #include "Vtop.h"

+using namespace sc_core;
+
 int sc_main(int argc, char* argv[]) {
    // See a similar example walkthrough in the verilator manpage.

@@ -31,7 +31,7 @@ CPPFLAGS += -DVL_DEBUG=1
 # Turn on some more compiler lint flags (when configured appropriately)
 # For testing inside Verilator, "configure --enable-ccwarn" will do this
 # automatically; otherwise you may want this unconditionally enabled
-ifeq ($(CFG_WITH_CCWARN),yes)	# Local... Else don't burden users
+ifeq ($(CFG_WITH_CCWARN),yes)  # Local... Else don't burden users
 USER_CPPFLAGS_WALL += -W -Werror -Wall
 endif

@@ -30,7 +30,7 @@ CPPFLAGS += -Wno-deprecated
 # Turn on some more flags (when configured appropriately)
 # For testing inside Verilator, "configure --enable-ccwarn" will do this
 # automatically; otherwise you may want this unconditionally enabled
-ifeq ($(CFG_WITH_CCWARN),yes)	# Local... Else don't burden users
+ifeq ($(CFG_WITH_CCWARN),yes)  # Local... Else don't burden users
 USER_CPPFLAGS_WALL += -W -Werror -Wall
 endif

@@ -10,7 +10,7 @@
 #include <memory>

 // SystemC global header
-#include <systemc.h>
+#include <systemc>

 // Include common routines
 #include <verilated.h>
@@ -23,6 +23,9 @@
 // Include model header, generated from Verilating "top.v"
 #include "Vtop.h"

+using namespace sc_core;
+using namespace sc_dt;
+
 int sc_main(int argc, char* argv[]) {
    // This is a more complicated example, please also see the simpler examples/make_hello_c.

@@ -50,7 +53,7 @@ int sc_main(int argc, char* argv[]) {
    Verilated::commandArgs(argc, argv);

    // General logfile
-    ios::sync_with_stdio();
+    std::ios::sync_with_stdio();

    // Define clocks
    sc_clock clk{"clk", 10, SC_NS, 0.5, 3, SC_NS, true};
@@ -60,8 +60,8 @@ class VlFileCopy:
            print("\t%s " % command)
        status = subprocess.call(command, shell=True)
        if status != 0:
-            raise Exception("Command failed running Verilator with '" +
-                            command + "', stopped")
+            raise RuntimeError("Command failed running Verilator with '" +
+                               command + "', stopped")


 #######################################################################
@@ -83,8 +83,8 @@ class VlHierGraph:
            print("\t%s " % command)
        status = subprocess.call(command, shell=True)
        if status != 0:
-            raise Exception("Command failed running Verilator with '" +
-                            command + "', stopped")
+            raise RuntimeError("Command failed running Verilator with '" +
+                               command + "', stopped")


 #######################################################################
@@ -58,6 +58,7 @@
 #endif

 #ifdef __MINGW32__
+#define WIN32_LEAN_AND_MEAN
 #include <windows.h>
 #endif

@@ -131,7 +132,7 @@ void **JenkinsIns(void *base_i, const unsigned char *mem, uint32_t length, uint3
 #define FST_GZIO_LEN                    (32768)
 #define FST_HDR_FOURPACK_DUO_SIZE       (4*1024*1024)

-#if defined(__i386__) || defined(__x86_64__) || defined(_AIX)
+#if defined(__i386__) || defined(__x86_64__) || defined(_AIX) || defined(__aarch64__)
 #define FST_DO_MISALIGNED_OPS
 #endif

@@ -334,26 +335,22 @@ return(NULL);
 /*
 * mmap compatibility
 */
-#if defined __CYGWIN__ || defined __MINGW32__
+#if defined __MINGW32__
 #include <limits.h>
 #define fstMmap(__addr,__len,__prot,__flags,__fd,__off) fstMmap2((__len), (__fd), (__off))
-#define fstMunmap(__addr,__len)                         free(__addr)
+#define fstMunmap(__addr,__len)                         UnmapViewOfFile((LPCVOID)__addr)

 static void *fstMmap2(size_t __len, int __fd, fst_off_t __off)
 {
-(void)__off;
+HANDLE handle = CreateFileMapping((HANDLE)_get_osfhandle(__fd), NULL,
+				  PAGE_READWRITE, (DWORD)(__len >> 32),
+				  (DWORD)__len, NULL);
+if (!handle) { return NULL; }

-unsigned char *pnt = (unsigned char *)malloc(__len);
-fst_off_t cur_offs = lseek(__fd, 0, SEEK_CUR);
-size_t i;
-
-lseek(__fd, 0, SEEK_SET);
-for(i=0;i<__len;i+=SSIZE_MAX)
-        {
-        read(__fd, pnt + i, ((__len - i) >= SSIZE_MAX) ? SSIZE_MAX : (__len - i));
-        }
-lseek(__fd, cur_offs, SEEK_SET);
-return(pnt);
+void *ptr = MapViewOfFileEx(handle, FILE_MAP_READ | FILE_MAP_WRITE,
+			    0, (DWORD)__off, (SIZE_T)__len, (LPVOID)NULL);
+CloseHandle(handle);
+return ptr;
 }
 #else
 #include <sys/mman.h>
@@ -984,14 +981,26 @@ fflush(xc->handle);
 */
 static void fstWriterMmapSanity(void *pnt, const char *file, int line, const char *usage)
 {
-#if !defined(__CYGWIN__) && !defined(__MINGW32__)
-if(pnt == MAP_FAILED)
+if(pnt == NULL
+#ifdef MAP_FAILED
+   || pnt == MAP_FAILED
+#endif
+  )
 	{
 	fprintf(stderr, "fstMmap() assigned to %s failed: errno: %d, file %s, line %d.\n", usage, errno, file, line);
+#if !defined(__MINGW32__)
 	perror("Why");
+#else
+	LPSTR mbuf = NULL;
+	FormatMessage(FORMAT_MESSAGE_ALLOCATE_BUFFER | FORMAT_MESSAGE_FROM_SYSTEM
+		      | FORMAT_MESSAGE_IGNORE_INSERTS, NULL, GetLastError(),
+		      MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT),
+		      (LPSTR)&mbuf, 0, NULL);
+	fprintf(stderr, "%s", mbuf);
+	LocalFree(mbuf);
+#endif
 	pnt = NULL;
 	}
-#endif
 }


@@ -1037,34 +1046,11 @@ if(!xc->curval_mem)

 static void fstDestroyMmaps(struct fstWriterContext *xc, int is_closing)
 {
-#if !defined __CYGWIN__ && !defined __MINGW32__
 (void)is_closing;
-#endif

 fstMunmap(xc->valpos_mem, xc->maxhandle * 4 * sizeof(uint32_t));
 xc->valpos_mem = NULL;

-#if defined __CYGWIN__ || defined __MINGW32__
-if(xc->curval_mem)
-        {
-        if(!is_closing) /* need to flush out for next emulated mmap() read */
-                {
-                unsigned char *pnt = xc->curval_mem;
-                int __fd = fileno(xc->curval_handle);
-                fst_off_t cur_offs = lseek(__fd, 0, SEEK_CUR);
-                size_t i;
-                size_t __len = xc->maxvalpos;
-
-                lseek(__fd, 0, SEEK_SET);
-                for(i=0;i<__len;i+=SSIZE_MAX)
-                        {
-                        write(__fd, pnt + i, ((__len - i) >= SSIZE_MAX) ? SSIZE_MAX : (__len - i));
-                        }
-                lseek(__fd, cur_offs, SEEK_SET);
-                }
-        }
-#endif
-
 fstMunmap(xc->curval_mem, xc->maxvalpos);
 xc->curval_mem = NULL;
 }
@@ -2066,7 +2052,8 @@ if(xc && !xc->already_in_close && !xc->already_in_flush)
                int zfd;
                int fourpack_duo = 0;
 #ifndef __MINGW32__
-                char *fnam = (char *)malloc(strlen(xc->filename) + 5 + 1);
+		int fnam_len = strlen(xc->filename) + 5 + 1;
+                char *fnam = (char *)malloc(fnam_len);
 #endif

                fixup_offs = ftello(xc->handle);
@@ -2156,7 +2143,7 @@ if(xc && !xc->already_in_close && !xc->already_in_flush)
                fflush(xc->handle);

 #ifndef __MINGW32__
-                sprintf(fnam, "%s.hier", xc->filename);
+                snprintf(fnam, fnam_len, "%s.hier", xc->filename);
                unlink(fnam);
                free(fnam);
 #endif
@@ -2835,7 +2822,7 @@ if(ctx && name && literal_arr && val_arr && (elem_count != 0))
 	uint32_t i;

 	name_len = strlen(name);
-	elem_count_len = sprintf(elem_count_buf, "%" PRIu32, elem_count);
+	elem_count_len = snprintf(elem_count_buf, 16, "%" PRIu32, elem_count);

 	literal_lens = (unsigned int *)calloc(elem_count, sizeof(unsigned int));
 	val_lens = (unsigned int *)calloc(elem_count, sizeof(unsigned int));
@@ -3412,6 +3399,7 @@ int flat_hier_alloc_len;
 unsigned do_rewind : 1;
 char str_scope_nam[FST_ID_NAM_SIZ+1];
 char str_scope_comp[FST_ID_NAM_SIZ+1];
+char *str_scope_attr;

 unsigned fseek_failed : 1;

@@ -3922,7 +3910,8 @@ int pass_status = 1;
 if(!xc->fh)
        {
        fst_off_t offs_cache = ftello(xc->f);
-        char *fnam = (char *)malloc(strlen(xc->filename) + 6 + 16 + 32 + 1);
+	int fnam_len = strlen(xc->filename) + 6 + 16 + 32 + 1;
+        char *fnam = (char *)malloc(fnam_len);
        unsigned char *mem = (unsigned char *)malloc(FST_GZIO_LEN);
        fst_off_t hl, uclen;
        fst_off_t clen = 0;
@@ -3941,7 +3930,7 @@ if(!xc->fh)
                htyp = xc->contains_hier_section_lz4duo ? FST_BL_HIER_LZ4DUO : FST_BL_HIER_LZ4;
                }

-        sprintf(fnam, "%s.hier_%d_%p", xc->filename, getpid(), (void *)xc);
+        snprintf(fnam, fnam_len, "%s.hier_%d_%p", xc->filename, getpid(), (void *)xc);
        fstReaderFseeko(xc, xc->f, xc->hier_pos, SEEK_SET);
        uclen = fstReaderUint64(xc->f);
 #ifndef __MINGW32__
@@ -4140,7 +4129,7 @@ if(!(isfeof=feof(xc->fh)))
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_SIZ)
+				if(cl < FST_ID_NAM_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
@@ -4152,7 +4141,7 @@ if(!(isfeof=feof(xc->fh)))
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_SIZ)
+				if(cl < FST_ID_NAM_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
@@ -4169,15 +4158,19 @@ if(!(isfeof=feof(xc->fh)))
                        xc->hier.htyp = FST_HT_ATTRBEGIN;
                        xc->hier.u.attr.typ = fgetc(xc->fh);
                        xc->hier.u.attr.subtype = fgetc(xc->fh);
-                        xc->hier.u.attr.name = pnt = xc->str_scope_nam;
+			if(!xc->str_scope_attr)
+				{
+				xc->str_scope_attr = (char *)calloc(1, FST_ID_NAM_ATTR_SIZ+1);
+				}
+                        xc->hier.u.attr.name = pnt = xc->str_scope_attr;
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_SIZ)
+				if(cl < FST_ID_NAM_ATTR_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
-                                }; /* scopename */
+                                }; /* attrname */
 			pnt[cl] = 0;
                        xc->hier.u.attr.name_length = cl;

@@ -4188,7 +4181,7 @@ if(!(isfeof=feof(xc->fh)))
                                if((xc->hier.u.attr.subtype == FST_MT_SOURCESTEM)||(xc->hier.u.attr.subtype == FST_MT_SOURCEISTEM))
                                        {
                                        int sidx_skiplen_dummy = 0;
-                                        xc->hier.u.attr.arg_from_name = fstGetVarint64((unsigned char *)xc->str_scope_nam, &sidx_skiplen_dummy);
+                                        xc->hier.u.attr.arg_from_name = fstGetVarint64((unsigned char *)xc->str_scope_attr, &sidx_skiplen_dummy);
                                        }
                                }
                        break;
@@ -4237,7 +4230,7 @@ if(!(isfeof=feof(xc->fh)))
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_SIZ)
+				if(cl < FST_ID_NAM_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
@@ -4376,7 +4369,7 @@ while(!feof(xc->fh))
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_ATTR_SIZ)
+				if(cl < FST_ID_NAM_ATTR_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
@@ -4398,7 +4391,7 @@ while(!feof(xc->fh))
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_ATTR_SIZ)
+				if(cl < FST_ID_NAM_ATTR_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
@@ -4487,7 +4480,7 @@ while(!feof(xc->fh))
 			cl = 0;
                        while((ch = fgetc(xc->fh)))
                                {
-				if(cl <= FST_ID_NAM_ATTR_SIZ)
+				if(cl < FST_ID_NAM_ATTR_SIZ)
 					{
 					pnt[cl++] = ch;
 					}
@@ -4594,15 +4587,17 @@ if(sectype == FST_BL_ZWRAPPER)
        int zfd;
        int flen = strlen(xc->filename);
        char *hf;
+	int hf_len;

        seclen = fstReaderUint64(xc->f);
        uclen = fstReaderUint64(xc->f);

        if(!seclen) return(0); /* not finished compressing, this is a failed read */

-        hf = (char *)calloc(1, flen + 16 + 32 + 1);
+	hf_len = flen + 16 + 32 + 1;
+        hf = (char *)calloc(1, hf_len);

-        sprintf(hf, "%s.upk_%d_%p", xc->filename, getpid(), (void *)xc);
+        snprintf(hf, hf_len, "%s.upk_%d_%p", xc->filename, getpid(), (void *)xc);
        fcomp = fopen(hf, "w+b");
        if(!fcomp)
                {
@@ -4974,6 +4969,7 @@ if(xc)
        free(xc->signal_typs); xc->signal_typs = NULL;
        free(xc->signal_lens); xc->signal_lens = NULL;
        free(xc->filename); xc->filename = NULL;
+	free(xc->str_scope_attr); xc->str_scope_attr = NULL;

        if(xc->fh)
                {
@@ -5197,16 +5193,16 @@ for(;;)

                                if(beg_tim)
                                        {
-					if(dumpvars_state == 1) { wx_len = sprintf(wx_buf, "$end\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 2; }
-                                        wx_len = sprintf(wx_buf, "#%" PRIu64 "\n", beg_tim);
+					if(dumpvars_state == 1) { wx_len = snprintf(wx_buf, 32, "$end\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 2; }
+                                        wx_len = snprintf(wx_buf, 32, "#%" PRIu64 "\n", beg_tim);
                                        fstWritex(xc, wx_buf, wx_len);
-					if(!dumpvars_state) { wx_len = sprintf(wx_buf, "$dumpvars\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 1; }
+					if(!dumpvars_state) { wx_len = snprintf(wx_buf, 32, "$dumpvars\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 1; }
                                        }
                                if((xc->num_blackouts)&&(cur_blackout != xc->num_blackouts))
                                        {
                                        if(beg_tim == xc->blackout_times[cur_blackout])
                                                {
-                                                wx_len = sprintf(wx_buf, "$dump%s $end\n", (xc->blackout_activity[cur_blackout++]) ? "on" : "off");
+                                                wx_len = snprintf(wx_buf, 32, "$dump%s $end\n", (xc->blackout_activity[cur_blackout++]) ? "on" : "off");
                                                fstWritex(xc, wx_buf, wx_len);
                                                }
                                        }
@@ -5340,7 +5336,7 @@ for(;;)
                                                                                        clone_d[j] = srcdata[7-j];
                                                                                        }
                                                                                }
-                                                                        sprintf((char *)xc->temp_signal_value_buf, "%.16g", d);
+                                                                        snprintf((char *)xc->temp_signal_value_buf, xc->longest_signal_value_len + 1, "%.16g", d);
                                                                        value_change_callback(user_callback_data_pointer, beg_tim, idx+1, xc->temp_signal_value_buf);
                                                                        }
                                                                }
@@ -5368,7 +5364,7 @@ for(;;)
                                                                                }

                                                                        fstVcdID(vcdid_buf, idx+1);
-                                                                        wx_len = sprintf(wx_buf, "r%.16g %s\n", d, vcdid_buf);
+                                                                        wx_len = snprintf(wx_buf, 64, "r%.16g %s\n", d, vcdid_buf);
                                                                        fstWritex(xc, wx_buf, wx_len);
                                                                        }
                                                                }
@@ -5633,16 +5629,16 @@ for(;;)
                                                }
                                        }

-				if(dumpvars_state == 1) { wx_len = sprintf(wx_buf, "$end\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 2; }
-                                wx_len = sprintf(wx_buf, "#%" PRIu64 "\n", time_table[i]);
+				if(dumpvars_state == 1) { wx_len = snprintf(wx_buf, 32, "$end\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 2; }
+                                wx_len = snprintf(wx_buf, 32, "#%" PRIu64 "\n", time_table[i]);
                                fstWritex(xc, wx_buf, wx_len);
-				if(!dumpvars_state) { wx_len = sprintf(wx_buf, "$dumpvars\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 1; }
+				if(!dumpvars_state) { wx_len = snprintf(wx_buf, 32, "$dumpvars\n"); fstWritex(xc, wx_buf, wx_len); dumpvars_state = 1; }

                                if((xc->num_blackouts)&&(cur_blackout != xc->num_blackouts))
                                        {
                                        if(time_table[i] == xc->blackout_times[cur_blackout])
                                                {
-                                                wx_len = sprintf(wx_buf, "$dump%s $end\n", (xc->blackout_activity[cur_blackout++]) ? "on" : "off");
+                                                wx_len = snprintf(wx_buf, 32, "$dump%s $end\n", (xc->blackout_activity[cur_blackout++]) ? "on" : "off");
                                                fstWritex(xc, wx_buf, wx_len);
                                                }
                                        }
@@ -5892,7 +5888,7 @@ for(;;)
                                                                        clone_d[j] = srcdata[7-j];
                                                                        }
                                                                }
-                                                        sprintf((char *)xc->temp_signal_value_buf, "%.16g", d);
+                                                        snprintf((char *)xc->temp_signal_value_buf, xc->longest_signal_value_len + 1, "%.16g", d);
                                                        value_change_callback(user_callback_data_pointer, time_table[i], idx+1, xc->temp_signal_value_buf);
                                                        }
                                                }
@@ -5918,7 +5914,7 @@ for(;;)
                                                                        }
                                                                }

-                                                        wx_len = sprintf(wx_buf, "r%.16g", d);
+                                                        wx_len = snprintf(wx_buf, 32, "r%.16g", d);
                                                        fstWritex(xc, wx_buf, wx_len);
                                                        }
                                                }
@@ -6024,7 +6020,7 @@ if(xc->signal_lens[facidx] == 1)
                                }
                        }

-                sprintf((char *)buf, "%.16g", d);
+                snprintf((char *)buf, 32, "%.16g", d); /* this will write 18 bytes */
                }
        }

@@ -6616,7 +6612,7 @@ if(xc->signal_lens[facidx] == 1)
                                        }
                                }

-                        sprintf(buf, "r%.16g", d);
+                        snprintf(buf, 32, "r%.16g", d); /* this will write 19 bytes */
                        return(buf);
                        }
                }
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -107,16 +107,9 @@ thread_local Verilated::ThreadLocal Verilated::t_s;

 #ifndef VL_USER_FINISH  ///< Define this to override the vl_finish function
 void vl_finish(const char* filename, int linenum, const char* hier) VL_MT_UNSAFE {
-    if (false && hier) {}
+    if (false && hier) {}  // Unused argument
    VL_PRINTF(  // Not VL_PRINTF_MT, already on main thread
        "- %s:%d: Verilog $finish\n", filename, linenum);
-    if (Verilated::threadContextp()->gotFinish()) {
-        VL_PRINTF(  // Not VL_PRINTF_MT, already on main thread
-            "- %s:%d: Second verilog $finish, exiting\n", filename, linenum);
-        Verilated::runFlushCallbacks();
-        Verilated::runExitCallbacks();
-        std::exit(0);
-    }
    Verilated::threadContextp()->gotFinish(true);
 }
 #endif
@@ -232,9 +225,9 @@ std::string _vl_string_vprintf(const char* formatp, va_list ap) VL_MT_SAFE {
    char* const bufp = new char[len + 1];
    VL_VSNPRINTF(bufp, len + 1, formatp, ap);

-    std::string out{bufp, len};  // Not const to allow move optimization
+    std::string result{bufp, len};  // Not const to allow move optimization
    delete[] bufp;
-    return out;
+    return result;
 }

 uint64_t _vl_dbg_sequence_number() VL_MT_SAFE {
@@ -255,30 +248,89 @@ void VL_DBG_MSGF(const char* formatp, ...) VL_MT_SAFE {
    // includes that otherwise would be required in every Verilated module
    va_list ap;
    va_start(ap, formatp);
-    const std::string out = _vl_string_vprintf(formatp, ap);
+    const std::string result = _vl_string_vprintf(formatp, ap);
    va_end(ap);
    // printf("-imm-V{t%d,%" PRId64 "}%s", VL_THREAD_ID(), _vl_dbg_sequence_number(),
-    // out.c_str());
+    // result.c_str());

    // Using VL_PRINTF not VL_PRINTF_MT so that we can call VL_DBG_MSGF
    // from within the guts of the thread execution machinery (and it goes
    // to the screen and not into the queues we're debugging)
-    VL_PRINTF("-V{t%u,%" PRIu64 "}%s", VL_THREAD_ID(), _vl_dbg_sequence_number(), out.c_str());
+    VL_PRINTF("-V{t%u,%" PRIu64 "}%s", VL_THREAD_ID(), _vl_dbg_sequence_number(), result.c_str());
 }

 void VL_PRINTF_MT(const char* formatp, ...) VL_MT_SAFE {
    va_list ap;
    va_start(ap, formatp);
-    const std::string out = _vl_string_vprintf(formatp, ap);
+    const std::string result = _vl_string_vprintf(formatp, ap);
    va_end(ap);
    VerilatedThreadMsgQueue::post(VerilatedMsg{[=]() {  //
-        VL_PRINTF("%s", out.c_str());
+        VL_PRINTF("%s", result.c_str());
    }});
 }

 //===========================================================================
 // Random -- Mostly called at init time, so not inline.

+VlRNG::VlRNG() VL_MT_SAFE {
+    // Starting point for this new class comes from the global RNG
+    VlRNG& fromr = vl_thread_rng();
+    m_state = fromr.m_state;
+    // Advance the *source* so it can later generate a new number
+    // Xoroshiro128+ algorithm
+    fromr.m_state[1] ^= fromr.m_state[0];
+    fromr.m_state[0] = (((fromr.m_state[0] << 55) | (fromr.m_state[0] >> 9)) ^ fromr.m_state[1]
+                        ^ (fromr.m_state[1] << 14));
+    fromr.m_state[1] = (fromr.m_state[1] << 36) | (fromr.m_state[1] >> 28);
+}
+uint64_t VlRNG::rand64() VL_MT_UNSAFE {
+    // Xoroshiro128+ algorithm
+    const uint64_t result = m_state[0] + m_state[1];
+    m_state[1] ^= m_state[0];
+    m_state[0] = (((m_state[0] << 55) | (m_state[0] >> 9)) ^ m_state[1] ^ (m_state[1] << 14));
+    m_state[1] = (m_state[1] << 36) | (m_state[1] >> 28);
+    return result;
+}
+uint64_t VlRNG::vl_thread_rng_rand64() VL_MT_SAFE {
+    VlRNG& fromr = vl_thread_rng();
+    const uint64_t result = fromr.m_state[0] + fromr.m_state[1];
+    fromr.m_state[1] ^= fromr.m_state[0];
+    fromr.m_state[0] = (((fromr.m_state[0] << 55) | (fromr.m_state[0] >> 9)) ^ fromr.m_state[1]
+                        ^ (fromr.m_state[1] << 14));
+    fromr.m_state[1] = (fromr.m_state[1] << 36) | (fromr.m_state[1] >> 28);
+    return result;
+}
+void VlRNG::srandom(uint64_t n) VL_MT_UNSAFE {
+    m_state[0] = n;
+    m_state[1] = m_state[0];
+    // Fix state as algorithm is slow to randomize if many zeros
+    // This causes a loss of ~ 1 bit of seed entropy, no big deal
+    if (VL_COUNTONES_I(m_state[0]) < 10) m_state[0] = ~m_state[0];
+    if (VL_COUNTONES_I(m_state[1]) < 10) m_state[1] = ~m_state[1];
+}
+std::string VlRNG::get_randstate() const VL_MT_UNSAFE {
+    // Though not stated in IEEE, assumption is the string must be printable
+    const char* const stateCharsp = reinterpret_cast<const char*>(&m_state);
+    static_assert(sizeof(m_state) == 16, "");
+    std::string result{"R00112233445566770011223344556677"};
+    for (size_t i = 0; i < sizeof(m_state); ++i) {
+        result[1 + i * 2] = 'a' + ((stateCharsp[i] >> 4) & 15);
+        result[1 + i * 2 + 1] = 'a' + (stateCharsp[i] & 15);
+    }
+    return result;
+}
+void VlRNG::set_randstate(const std::string& state) VL_MT_UNSAFE {
+    if (VL_UNLIKELY((state.length() != 1 + 2 * sizeof(m_state)) || (state[0] != 'R'))) {
+        VL_PRINTF_MT("%%Warning: set_randstate ignored as state string not from get_randstate\n");
+        return;
+    }
+    char* const stateCharsp = reinterpret_cast<char*>(&m_state);
+    for (size_t i = 0; i < sizeof(m_state); ++i) {
+        stateCharsp[i]
+            = (((state[1 + i * 2] - 'a') & 15) << 4) | ((state[1 + i * 2 + 1] - 'a') & 15);
+    }
+}
+
 static uint32_t vl_sys_rand32() VL_MT_SAFE {
    // Return random 32-bits using system library.
    // Used only to construct seed for Verilator's PRNG.
@@ -292,27 +344,23 @@ static uint32_t vl_sys_rand32() VL_MT_SAFE {
 #endif
 }

-uint64_t vl_rand64() VL_MT_SAFE {
-    static thread_local uint64_t t_state[2];
+VlRNG& VlRNG::vl_thread_rng() VL_MT_SAFE {
+    static thread_local VlRNG t_rng{0};
    static thread_local uint32_t t_seedEpoch = 0;
    // For speed, we use a thread-local epoch number to know when to reseed
    // A thread always belongs to a single context, so this works out ok
    if (VL_UNLIKELY(t_seedEpoch != VerilatedContextImp::randSeedEpoch())) {
        // Set epoch before state, to avoid race case with new seeding
        t_seedEpoch = VerilatedContextImp::randSeedEpoch();
-        t_state[0] = Verilated::threadContextp()->impp()->randSeedDefault64();
-        t_state[1] = t_state[0];
+        // Same as srandom() but here as needs to be VL_MT_SAFE
+        t_rng.m_state[0] = Verilated::threadContextp()->impp()->randSeedDefault64();
+        t_rng.m_state[1] = t_rng.m_state[0];
        // Fix state as algorithm is slow to randomize if many zeros
        // This causes a loss of ~ 1 bit of seed entropy, no big deal
-        if (VL_COUNTONES_I(t_state[0]) < 10) t_state[0] = ~t_state[0];
-        if (VL_COUNTONES_I(t_state[1]) < 10) t_state[1] = ~t_state[1];
+        if (VL_COUNTONES_I(t_rng.m_state[0]) < 10) t_rng.m_state[0] = ~t_rng.m_state[0];
+        if (VL_COUNTONES_I(t_rng.m_state[1]) < 10) t_rng.m_state[1] = ~t_rng.m_state[1];
    }
-    // Xoroshiro128+ algorithm
-    const uint64_t result = t_state[0] + t_state[1];
-    t_state[1] ^= t_state[0];
-    t_state[0] = (((t_state[0] << 55) | (t_state[0] >> 9)) ^ t_state[1] ^ (t_state[1] << 14));
-    t_state[1] = (t_state[1] << 36) | (t_state[1] >> 28);
-    return result;
+    return t_rng;
 }

 WDataOutP VL_RANDOM_W(int obits, WDataOutP outwp) VL_MT_SAFE {
@@ -321,6 +369,12 @@ WDataOutP VL_RANDOM_W(int obits, WDataOutP outwp) VL_MT_SAFE {
    return outwp;
 }

+WDataOutP VL_RANDOM_RNG_W(VlRNG& rngr, int obits, WDataOutP outwp) VL_MT_UNSAFE {
+    for (int i = 0; i < VL_WORDS_I(obits); ++i) outwp[i] = rngr.rand64();
+    // Last word is unclean
+    return outwp;
+}
+
 IData VL_RANDOM_SEEDED_II(IData& seedr) VL_MT_SAFE {
    // $random - seed is a new seed to apply, then we return new seed
    Verilated::threadContextp()->randSeed(static_cast<int>(seedr));
@@ -390,6 +444,8 @@ WDataOutP _vl_moddiv_w(int lbits, WDataOutP owp, const WDataInP lwp, const WData

    const int uw = VL_WORDS_I(umsbp1);  // aka "m" in the algorithm
    const int vw = VL_WORDS_I(vmsbp1);  // aka "n" in the algorithm
+    VL_DEBUG_IFDEF(assert(uw <= VL_MULS_MAX_WORDS););
+    VL_DEBUG_IFDEF(assert(vw <= VL_MULS_MAX_WORDS););

    if (vw == 1) {  // Single divisor word breaks rest of algorithm
        uint64_t k = 0;
@@ -411,27 +467,25 @@ WDataOutP _vl_moddiv_w(int lbits, WDataOutP owp, const WDataInP lwp, const WData

    // Zero for ease of debugging and to save having to zero for shifts
    // Note +1 as loop will use extra word
-    for (int i = 0; i < words + 1; ++i) { un[i] = vn[i] = 0; }
+    for (int i = 0; i < words + 1; ++i) un[i] = vn[i] = 0;

    // Algorithm requires divisor MSB to be set
    // Copy and shift to normalize divisor so MSB of vn[vw-1] is set
    const int s = 31 - VL_BITBIT_I(vmsbp1 - 1);  // shift amount (0...31)
-    const uint32_t shift_mask = s ? 0xffffffff : 0;  // otherwise >> 32 won't mask the value
-    for (int i = vw - 1; i > 0; --i) {
-        vn[i] = (rwp[i] << s) | (shift_mask & (rwp[i - 1] >> (32 - s)));
-    }
-    vn[0] = rwp[0] << s;
-
    // Copy and shift dividend by same amount; may set new upper word
    if (s) {
+        for (int i = vw - 1; i > 0; --i) vn[i] = (rwp[i] << s) | (rwp[i - 1] >> (32 - s));
+        vn[0] = rwp[0] << s;
        un[uw] = lwp[uw - 1] >> (32 - s);
+        for (int i = uw - 1; i > 0; --i) un[i] = (lwp[i] << s) | (lwp[i - 1] >> (32 - s));
+        un[0] = lwp[0] << s;
    } else {
+        for (int i = vw - 1; i > 0; --i) vn[i] = rwp[i];
+        vn[0] = rwp[0];
        un[uw] = 0;
+        for (int i = uw - 1; i > 0; --i) un[i] = lwp[i];
+        un[0] = lwp[0];
    }
-    for (int i = uw - 1; i > 0; --i) {
-        un[i] = (lwp[i] << s) | (shift_mask & (lwp[i - 1] >> (32 - s)));
-    }
-    un[0] = lwp[0] << s;

    // Main loop
    for (int j = uw - vw; j >= 0; --j) {
@@ -475,8 +529,10 @@ WDataOutP _vl_moddiv_w(int lbits, WDataOutP owp, const WDataInP lwp, const WData

    if (is_modulus) {  // modulus
        // Need to reverse normalization on copy to output
-        for (int i = 0; i < vw; ++i) {
-            owp[i] = (un[i] >> s) | (shift_mask & (un[i + 1] << (32 - s)));
+        if (s) {
+            for (int i = 0; i < vw; ++i) owp[i] = (un[i] >> s) | (un[i + 1] << (32 - s));
+        } else {
+            for (int i = 0; i < vw; ++i) owp[i] = un[i];
        }
        for (int i = vw; i < words; ++i) owp[i] = 0;
        return owp;
@@ -488,6 +544,8 @@ WDataOutP _vl_moddiv_w(int lbits, WDataOutP owp, const WDataInP lwp, const WData
 WDataOutP VL_POW_WWW(int obits, int, int rbits, WDataOutP owp, const WDataInP lwp,
                     const WDataInP rwp) VL_MT_SAFE {
    // obits==lbits, rbits can be different
+    const int owords = VL_WORDS_I(obits);
+    VL_DEBUG_IFDEF(assert(owords <= VL_MULS_MAX_WORDS););
    owp[0] = 1;
    for (int i = 1; i < VL_WORDS_I(obits); i++) owp[i] = 0;
    // cppcheck-has-bug-suppress variableScope
@@ -499,11 +557,11 @@ WDataOutP VL_POW_WWW(int obits, int, int rbits, WDataOutP owp, const WDataInP lw
    for (int bit = 0; bit < rbits; bit++) {
        if (bit > 0) {  // power = power*power
            VL_ASSIGN_W(obits, lastpowstore, powstore);
-            VL_MUL_W(VL_WORDS_I(obits), powstore, lastpowstore, lastpowstore);
+            VL_MUL_W(owords, powstore, lastpowstore, lastpowstore);
        }
        if (VL_BITISSET_W(rwp, bit)) {  // out *= power
            VL_ASSIGN_W(obits, lastoutstore, owp);
-            VL_MUL_W(VL_WORDS_I(obits), owp, lastoutstore, powstore);
+            VL_MUL_W(owords, owp, lastoutstore, powstore);
        }
    }
    return owp;
@@ -515,15 +573,18 @@ WDataOutP VL_POW_WWQ(int obits, int lbits, int rbits, WDataOutP owp, const WData
    return VL_POW_WWW(obits, lbits, rbits, owp, lwp, rhsw);
 }
 QData VL_POW_QQW(int, int, int rbits, QData lhs, const WDataInP rwp) VL_MT_SAFE {
-    // Skip check for rhs == 0, as short-circuit doesn't save time
+    const int rwords = VL_WORDS_I(rbits);
+    EData rnz = rwp[0];
+    for (int w = 1; w < rwords; ++w) rnz |= rwp[w];
+    if (!rnz) return 1;  // rwp == 0
    if (VL_UNLIKELY(lhs == 0)) return 0;
    QData power = lhs;
-    QData out = 1ULL;
+    QData result = 1ULL;
    for (int bit = 0; bit < rbits; ++bit) {
        if (bit > 0) power = power * power;
-        if (VL_BITISSET_W(rwp, bit)) out *= power;
+        if (VL_BITISSET_W(rwp, bit)) result *= power;
    }
-    return out;
+    return result;
 }

 WDataOutP VL_POWSS_WWW(int obits, int, int rbits, WDataOutP owp, const WDataInP lwp,
@@ -596,8 +657,10 @@ double VL_ITOR_D_W(int lbits, const WDataInP lwp) VL_PURE {
 }
 double VL_ISTOR_D_W(int lbits, const WDataInP lwp) VL_MT_SAFE {
    if (!VL_SIGN_W(lbits, lwp)) return VL_ITOR_D_W(lbits, lwp);
+    const int words = VL_WORDS_I(lbits);
+    VL_DEBUG_IFDEF(assert(words <= VL_MULS_MAX_WORDS););
    uint32_t pos[VL_MULS_MAX_WORDS + 1];  // Fixed size, as MSVC++ doesn't allow [words] here
-    VL_NEGATE_W(VL_WORDS_I(lbits), pos, lwp);
+    VL_NEGATE_W(words, pos, lwp);
    _vl_clean_inplace_w(lbits, pos);
    return -VL_ITOR_D_W(lbits, pos);
 }
@@ -645,7 +708,7 @@ std::string VL_DECIMAL_NW(int width, const WDataInP lwp) VL_MT_SAFE {
 }

 template <typename T>
-std::string _vl_vsformat_time(char* tmp, T ld, int timeunit, bool left, size_t width) {
+std::string _vl_vsformat_time(char* tmp, T ld, int timeunit, bool left, size_t width) VL_MT_SAFE {
    const VerilatedContextImp* const ctxImpp = Verilated::threadContextp()->impp();
    const std::string suffix = ctxImpp->timeFormatSuffix();
    const int userUnits = ctxImpp->timeFormatUnits();  // 0..-15
@@ -1312,16 +1375,18 @@ IData _vl_vsscanf(FILE* fp,  // If a fscanf
                    *p = t_tmp;
                } else if (obits <= VL_BYTESIZE) {
                    CData* const p = va_arg(ap, CData*);
-                    *p = owp[0];
+                    *p = VL_CLEAN_II(obits, obits, owp[0]);
                } else if (obits <= VL_SHORTSIZE) {
                    SData* const p = va_arg(ap, SData*);
-                    *p = owp[0];
+                    *p = VL_CLEAN_II(obits, obits, owp[0]);
                } else if (obits <= VL_IDATASIZE) {
                    IData* const p = va_arg(ap, IData*);
-                    *p = owp[0];
+                    *p = VL_CLEAN_II(obits, obits, owp[0]);
                } else if (obits <= VL_QUADSIZE) {
                    QData* const p = va_arg(ap, QData*);
-                    *p = VL_SET_QW(owp);
+                    *p = VL_CLEAN_QQ(obits, obits, VL_SET_QW(owp));
+                } else {
+                    _vl_clean_inplace_w(obits, owp);
                }
            }
            }  // switch
@@ -1411,6 +1476,12 @@ IData VL_FERROR_IN(IData, std::string& outputr) VL_MT_SAFE {
    outputr = std::string{::std::strerror(ret)};
    return ret;
 }
+IData VL_FERROR_IW(IData fpi, int obits, WDataOutP outwp) VL_MT_SAFE {
+    std::string output;
+    const IData ret = VL_FERROR_IN(fpi, output /*ref*/);
+    _vl_string_to_vint(obits, outwp, output.length(), output.c_str());
+    return ret;
+}

 IData VL_FOPEN_NN(const std::string& filename, const std::string& mode) {
    return Verilated::threadContextp()->impp()->fdNew(filename.c_str(), mode.c_str());
@@ -1645,15 +1716,15 @@ std::string VL_STACKTRACE_N() VL_MT_SAFE {
    // cppcheck-suppress knownConditionTrueFalse
    if (!strings) return "Unable to backtrace\n";

-    std::string out = "Backtrace:\n";
-    for (int j = 0; j < nptrs; j++) out += std::string{strings[j]} + std::string{"\n"};
+    std::string result = "Backtrace:\n";
+    for (int j = 0; j < nptrs; j++) result += std::string{strings[j]} + std::string{"\n"};
    free(strings);
-    return out;
+    return result;
 }

 void VL_STACKTRACE() VL_MT_SAFE {
-    const std::string out = VL_STACKTRACE_N();
-    VL_PRINTF("%s", out.c_str());
+    const std::string result = VL_STACKTRACE_N();
+    VL_PRINTF("%s", result.c_str());
 }

 IData VL_SYSTEM_IQ(QData lhs) VL_MT_SAFE {
@@ -1793,19 +1864,20 @@ std::string VL_TO_STRING(CData lhs) { return VL_SFORMATF_NX("'h%0x", 8, lhs); }
 std::string VL_TO_STRING(SData lhs) { return VL_SFORMATF_NX("'h%0x", 16, lhs); }
 std::string VL_TO_STRING(IData lhs) { return VL_SFORMATF_NX("'h%0x", 32, lhs); }
 std::string VL_TO_STRING(QData lhs) { return VL_SFORMATF_NX("'h%0x", 64, lhs); }
+std::string VL_TO_STRING(double lhs) { return VL_SFORMATF_NX("%d", 64, lhs); }
 std::string VL_TO_STRING_W(int words, const WDataInP obj) {
    return VL_SFORMATF_NX("'h%0x", words * VL_EDATASIZE, obj);
 }

 std::string VL_TOLOWER_NN(const std::string& ld) VL_PURE {
-    std::string out = ld;
-    for (auto& cr : out) cr = std::tolower(cr);
-    return out;
+    std::string result = ld;
+    for (auto& cr : result) cr = std::tolower(cr);
+    return result;
 }
 std::string VL_TOUPPER_NN(const std::string& ld) VL_PURE {
-    std::string out = ld;
-    for (auto& cr : out) cr = std::toupper(cr);
-    return out;
+    std::string result = ld;
+    for (auto& cr : result) cr = std::toupper(cr);
+    return result;
 }

 std::string VL_CVT_PACK_STR_NW(int lwords, const WDataInP lwp) VL_PURE {
@@ -1826,6 +1898,12 @@ std::string VL_CVT_PACK_STR_NW(int lwords, const WDataInP lwp) VL_PURE {
    return std::string{destout, len};
 }

+std::string VL_CVT_PACK_STR_ND(const VlQueue<std::string>& q) VL_PURE {
+    std::string output;
+    for (const std::string& s : q) output += s;
+    return output;
+}
+
 std::string VL_PUTC_N(const std::string& lhs, IData rhs, CData ths) VL_PURE {
    std::string lstring = lhs;
    const int32_t rhs_s = rhs;  // To signed value
@@ -1861,6 +1939,32 @@ IData VL_ATOI_N(const std::string& str, int base) VL_PURE {
    if (errno != 0) v = 0;
    return static_cast<IData>(v);
 }
+IData VL_NTOI_I(int obits, const std::string& str) VL_PURE { return VL_NTOI_Q(obits, str); }
+QData VL_NTOI_Q(int obits, const std::string& str) VL_PURE {
+    QData out = 0;
+    const char* const datap = str.data();
+    int pos = static_cast<int>(str.length()) - 1;
+    int bit = 0;
+    while (bit < obits && pos >= 0) {
+        out |= static_cast<QData>(datap[pos]) << VL_BITBIT_Q(bit);
+        bit += 8;
+        --pos;
+    }
+    return out & VL_MASK_Q(obits);
+}
+void VL_NTOI_W(int obits, WDataOutP owp, const std::string& str) VL_PURE {
+    const int words = VL_WORDS_I(obits);
+    for (int i = 0; i < words; ++i) owp[i] = 0;
+    const char* const datap = str.data();
+    int pos = static_cast<int>(str.length()) - 1;
+    int bit = 0;
+    while (bit < obits && pos >= 0) {
+        owp[VL_BITWORD_I(bit)] |= static_cast<EData>(datap[pos]) << VL_BITBIT_I(bit);
+        bit += 8;
+        --pos;
+    }
+    owp[words - 1] &= VL_MASK_E(obits);
+}

 //===========================================================================
 // Readmem/writemem
@@ -1992,7 +2096,7 @@ bool VlReadMem::get(QData& addrr, std::string& valuer) {

    if (VL_UNLIKELY(m_end != ~0ULL && m_addr <= m_end && !m_anyAddr)) {
        VL_WARN_MT(m_filename.c_str(), m_linenum, "",
-                   "$readmem file ended before specified final address (IEEE 2017 21.4)");
+                   "$readmem file ended before specified final address (IEEE 1800-2017 21.4)");
    }

    return false;  // EOF
@@ -2506,6 +2610,14 @@ VerilatedVirtualBase* VerilatedContext::threadPoolp() {
    return m_threadPool.get();
 }

+void VerilatedContext::prepareClone() { delete m_threadPool.release(); }
+
+VerilatedVirtualBase* VerilatedContext::threadPoolpOnClone() {
+    if (VL_UNLIKELY(m_threadPool)) m_threadPool.release();
+    m_threadPool = std::unique_ptr<VlThreadPool>(new VlThreadPool{this, m_threads - 1});
+    return m_threadPool.get();
+}
+
 VerilatedVirtualBase*
 VerilatedContext::enableExecutionProfiler(VerilatedVirtualBase* (*construct)(VerilatedContext&)) {
    if (!m_executionProfiler) m_executionProfiler.reset(construct(*this));
@@ -2600,14 +2712,11 @@ void VerilatedContextImp::commandArgVl(const std::string& arg) {
                        "Exiting due to command line argument (not an error)");
        } else if (arg == "+verilator+noassert") {
            assertOn(false);
-        } else if (commandArgVlUint64(arg, "+verilator+prof+exec+start+", u64)
-                   || commandArgVlUint64(arg, "+verilator+prof+threads+start+", u64)) {
+        } else if (commandArgVlUint64(arg, "+verilator+prof+exec+start+", u64)) {
            profExecStart(u64);
-        } else if (commandArgVlUint64(arg, "+verilator+prof+exec+window+", u64, 1)
-                   || commandArgVlUint64(arg, "+verilator+prof+threads+window+", u64, 1)) {
+        } else if (commandArgVlUint64(arg, "+verilator+prof+exec+window+", u64, 1)) {
            profExecWindow(u64);
-        } else if (commandArgVlString(arg, "+verilator+prof+exec+file+", str)
-                   || commandArgVlString(arg, "+verilator+prof+threads+file+", str)) {
+        } else if (commandArgVlString(arg, "+verilator+prof+exec+file+", str)) {
            profExecFilename(str);
        } else if (commandArgVlString(arg, "+verilator+prof+vlt+file+", str)) {
            profVltFilename(str);
@@ -2914,6 +3023,7 @@ void Verilated::endOfEval(VerilatedEvalMsgQueue* evalMsgQp) VL_MT_SAFE {
    // It should be ok to call into endOfEvalGuts, it returns immediately
    // if there are no transactions.
    VL_DEBUG_IF(VL_DBG_MSGF("End-of-eval cleanup\n"););
+    VerilatedThreadMsgQueue::flush(evalMsgQp);
    evalMsgQp->process();
 }

@@ -2948,7 +3058,7 @@ VerilatedModule::~VerilatedModule() {
 // VerilatedVar:: Methods

 // cppcheck-suppress unusedFunction  // Used by applications
-uint32_t VerilatedVarProps::entSize() const {
+uint32_t VerilatedVarProps::entSize() const VL_MT_SAFE {
    uint32_t size = 1;
    switch (vltype()) {
    case VLVT_PTR: size = sizeof(void*); break;
@@ -2968,7 +3078,7 @@ size_t VerilatedVarProps::totalSize() const {
    return size;
 }

-void* VerilatedVarProps::datapAdjustIndex(void* datap, int dim, int indx) const {
+void* VerilatedVarProps::datapAdjustIndex(void* datap, int dim, int indx) const VL_MT_SAFE {
    if (VL_UNLIKELY(dim <= 0 || dim > udims())) return nullptr;
    if (VL_UNLIKELY(indx < low(dim) || indx > high(dim))) return nullptr;
    const int indxAdj = indx - low(dim);
@@ -3087,7 +3197,7 @@ void* VerilatedScope::exportFindNullError(int funcnum) VL_MT_SAFE {
    return nullptr;
 }

-void* VerilatedScope::exportFindError(int funcnum) const {
+void* VerilatedScope::exportFindError(int funcnum) const VL_MT_SAFE {
    // Slowpath - Called only when find has failed
    const std::string msg
        = (std::string{"Testbench C called '"} + VerilatedImp::exportName(funcnum)
@@ -3132,15 +3242,18 @@ void VerilatedAssertOneThread::fatal_different() VL_MT_SAFE {
 //===========================================================================
 // VlDeleter:: Methods

-void VlDeleter::deleteAll() {
+void VlDeleter::deleteAll() VL_EXCLUDES(m_mutex) VL_EXCLUDES(m_deleteMutex) VL_MT_SAFE {
    while (true) {
-        VerilatedLockGuard lock{m_mutex};
-        if (m_newGarbage.empty()) break;
-        VerilatedLockGuard deleteLock{m_deleteMutex};
-        std::swap(m_newGarbage, m_toDelete);
-        lock.unlock();  // So destructors can enqueue new objects
-        for (VlDeletable* const objp : m_toDelete) delete objp;
-        m_toDelete.clear();
+        {
+            VerilatedLockGuard lock{m_mutex};
+            if (m_newGarbage.empty()) break;
+            m_deleteMutex.lock();
+            std::swap(m_newGarbage, m_deleteNow);
+            // m_mutex is unlocked here, so destructors can enqueue new objects
+        }
+        for (VlDeletable* const objp : m_deleteNow) delete objp;
+        m_deleteNow.clear();
+        m_deleteMutex.unlock();
    }
 }

@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -30,6 +30,13 @@
 #ifndef VERILATOR_VERILATED_H_
 #define VERILATOR_VERILATED_H_
 #define VERILATOR_VERILATED_H_INTERNAL_
+#ifdef VERILATOR_INTERNAL_
+// This file contains definition of VerilationMutex that should
+// only be used by verilated code. Verilator itself should use
+// mutex from V3Mutex.h. Make sure this file isn't included in
+// verilator code.
+#error "verilated.h should only be included in verilated code"
+#endif

 // clang-format off
 #include "verilatedos.h"
@@ -47,6 +54,7 @@
 #include <cstdlib>
 #include <cstring>
 #include <deque>
+#include <functional>
 #include <map>
 #include <memory>
 #include <set>
@@ -125,7 +133,8 @@ enum VerilatedVarType : uint8_t {
    VLVT_UINT32,  // AKA IData
    VLVT_UINT64,  // AKA QData
    VLVT_WDATA,  // AKA WData
-    VLVT_STRING  // C++ string
+    VLVT_STRING,  // C++ string
+    VLVT_REAL  // AKA double
 };

 enum VerilatedVarFlags {
@@ -141,13 +150,24 @@ enum VerilatedVarFlags {
    VLVF_DPI_CLAY = (1 << 10)  // DPI compatible C standard layout
 };

+//=============================================================================
+// Utility functions
+
+template <size_t N>
+inline constexpr size_t roundUpToMultipleOf(size_t value) {
+    static_assert((N & (N - 1)) == 0, "'N' must be a power of 2");
+    return (value + N - 1) & ~(N - 1);
+}
+
 //=========================================================================
 // Mutex and threading support

 // Return current thread ID (or 0), not super fast, cache if needed
 extern uint32_t VL_THREAD_ID() VL_MT_SAFE;

+#ifndef VL_LOCK_SPINS
 #define VL_LOCK_SPINS 50000  /// Number of times to spin for a mutex before yielding
+#endif

 /// Mutex, wrapped to allow -fthread_safety checks
 class VL_CAPABILITY("mutex") VerilatedMutex final {
@@ -158,6 +178,7 @@ public:
    /// Construct mutex (without locking it)
    VerilatedMutex() = default;
    ~VerilatedMutex() = default;
+    VL_UNCOPYABLE(VerilatedMutex);
    const VerilatedMutex& operator!() const { return *this; }  // For -fthread_safety
    /// Acquire/lock mutex
    void lock() VL_ACQUIRE() VL_MT_SAFE {
@@ -175,6 +196,19 @@ public:
    void unlock() VL_RELEASE() VL_MT_SAFE { m_mutex.unlock(); }
    /// Try to acquire mutex.  Returns true on success, and false on failure.
    bool try_lock() VL_TRY_ACQUIRE(true) VL_MT_SAFE { return m_mutex.try_lock(); }
+    /// Acquire/lock mutex and check for stop request
+    /// It tries to lock the mutex and if it fails, it check if stop request was send.
+    /// It returns after locking mutex.
+    /// This function should be extracted to V3ThreadPool, but due to clang thread-safety
+    /// limitations it needs to be placed here.
+    void lockCheckStopRequest(std::function<void()> checkStopRequestFunction)
+        VL_ACQUIRE() VL_MT_SAFE {
+        while (true) {
+            checkStopRequestFunction();
+            if (m_mutex.try_lock()) return;
+            VL_CPU_RELAX();
+        }
+    }
 };

 /// Lock guard for mutex (ala std::unique_lock), wrapped to allow -fthread_safety checks
@@ -188,14 +222,10 @@ public:
    /// Construct and hold given mutex lock until destruction or unlock()
    explicit VerilatedLockGuard(VerilatedMutex& mutexr) VL_ACQUIRE(mutexr) VL_MT_SAFE
        : m_mutexr(mutexr) {  // Need () or GCC 4.8 false warning
-        m_mutexr.lock();
+        mutexr.lock();
    }
    /// Destruct and unlock the mutex
    ~VerilatedLockGuard() VL_RELEASE() { m_mutexr.unlock(); }
-    /// Unlock the mutex
-    void lock() VL_ACQUIRE() VL_MT_SAFE { m_mutexr.lock(); }
-    /// Lock the mutex
-    void unlock() VL_RELEASE() VL_MT_SAFE { m_mutexr.unlock(); }
 };

 // Internals: Remember the calling thread at construction time, and make
@@ -541,6 +571,8 @@ public:
    void addModel(VerilatedModel*);

    VerilatedVirtualBase* threadPoolp();
+    void prepareClone();
+    VerilatedVirtualBase* threadPoolpOnClone();
    VerilatedVirtualBase*
    enableExecutionProfiler(VerilatedVirtualBase* (*construct)(VerilatedContext&));

@@ -584,6 +616,7 @@ public:  // But for internal use only
    VerilatedEvalMsgQueue* __Vm_evalMsgQp;
    explicit VerilatedSyms(VerilatedContext* contextp);  // Pass null for default context
    ~VerilatedSyms();
+    VL_UNCOPYABLE(VerilatedSyms);
 };

 //===========================================================================
@@ -594,8 +627,9 @@ class VerilatedScope final {
 public:
    enum Type : uint8_t {
        SCOPE_MODULE,
-        SCOPE_OTHER
-    };  // Type of a scope, currently module is only interesting
+        SCOPE_OTHER,
+        SCOPE_PACKAGE
+    };  // Type of a scope, currently only module and package are interesting
 private:
    // Fastpath:
    VerilatedSyms* m_symsp = nullptr;  // Symbol table
@@ -624,7 +658,7 @@ public:  // But internals only - called from VerilatedModule's
    VerilatedVar* varFind(const char* namep) const VL_MT_SAFE_POSTINIT;
    VerilatedVarNameMap* varsp() const VL_MT_SAFE_POSTINIT { return m_varsp; }
    void scopeDump() const;
-    void* exportFindError(int funcnum) const;
+    void* exportFindError(int funcnum) const VL_MT_SAFE;
    static void* exportFindNullError(int funcnum) VL_MT_SAFE;
    static void* exportFind(const VerilatedScope* scopep, int funcnum) VL_MT_SAFE {
        if (VL_UNLIKELY(!scopep)) return exportFindNullError(funcnum);
@@ -889,6 +923,7 @@ public:

    // Internal: Called at end of each thread mtask, before finishing eval
    static void endOfThreadMTask(VerilatedEvalMsgQueue* evalMsgQp) VL_MT_SAFE {
+        mtaskId(0);
        if (VL_UNLIKELY(t_s.t_endOfEvalReqd)) endOfThreadMTaskGuts(evalMsgQp);
    }
    // Internal: Called at end of eval loop
@@ -922,18 +957,18 @@ void VerilatedContext::timeprecision(int value) VL_MT_SAFE {
        const VerilatedLockGuard lock{m_mutex};
        m_s.m_timeprecision = value;
 #if VM_SC
-        const sc_time sc_res = sc_get_time_resolution();
-        if (sc_res == sc_time(1, SC_SEC)) {
+        const sc_core::sc_time sc_res = sc_core::sc_get_time_resolution();
+        if (sc_res == sc_core::sc_time(1, sc_core::SC_SEC)) {
            sc_prec = 0;
-        } else if (sc_res == sc_time(1, SC_MS)) {
+        } else if (sc_res == sc_core::sc_time(1, sc_core::SC_MS)) {
            sc_prec = 3;
-        } else if (sc_res == sc_time(1, SC_US)) {
+        } else if (sc_res == sc_core::sc_time(1, sc_core::SC_US)) {
            sc_prec = 6;
-        } else if (sc_res == sc_time(1, SC_NS)) {
+        } else if (sc_res == sc_core::sc_time(1, sc_core::SC_NS)) {
            sc_prec = 9;
-        } else if (sc_res == sc_time(1, SC_PS)) {
+        } else if (sc_res == sc_core::sc_time(1, sc_core::SC_PS)) {
            sc_prec = 12;
-        } else if (sc_res == sc_time(1, SC_FS)) {
+        } else if (sc_res == sc_core::sc_time(1, sc_core::SC_FS)) {
            sc_prec = 15;
        }
 #endif
@@ -2,13 +2,16 @@
 ######################################################################
 # DESCRIPTION: Makefile commands for all verilated target files
 #
-# Copyright 2003-2023 by Wilson Snyder. This program is free software; you
+# Copyright 2003-2024 by Wilson Snyder. This program is free software; you
 # can redistribute it and/or modify it under the terms of either the GNU
 # Lesser General Public License Version 3 or the Perl Artistic License
 # Version 2.0.
 # SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0
 ######################################################################

+# Tool names.
+# These are computed at configuration time, and most are not ?=
+# to avoid picking up potentially incorrect Make implicit variables
 AR = @AR@
 CXX = @CXX@
 LINK = @CXX@
@@ -16,6 +19,7 @@ OBJCACHE ?= @OBJCACHE@
 PERL = @PERL@
 PYTHON3 = @PYTHON3@

+# Configuration time options
 CFG_WITH_CCWARN = @CFG_WITH_CCWARN@
 CFG_WITH_LONGTESTS = @CFG_WITH_LONGTESTS@

@@ -31,6 +35,14 @@ CFG_CXXFLAGS_NO_UNUSED = @CFG_CXXFLAGS_NO_UNUSED@
 CFG_CXXFLAGS_WEXTRA = @CFG_CXXFLAGS_WEXTRA@
 # Compiler flags that enable coroutine support
 CFG_CXXFLAGS_COROUTINES = @CFG_CXXFLAGS_COROUTINES@
+# Compiler flags when creating a precompiled header
+CFG_CXXFLAGS_PCH = -x c++-header
+# Compiler option to put in front of filename to read precompiled header
+CFG_CXXFLAGS_PCH_I = @CFG_CXXFLAGS_PCH_I@
+# Compiler's filename prefix for precompiled headers, .gch if clang, empty if GCC
+CFG_GCH_IF_CLANG = @CFG_GCH_IF_CLANG@
+# Linker flags
+CFG_LDFLAGS_VERILATED = @CFG_LDFLAGS_VERILATED@
 # Linker libraries for multithreading
 CFG_LDLIBS_THREADS = @CFG_LDLIBS_THREADS@

@@ -41,6 +53,12 @@ VERILATOR_COVERAGE = $(PERL) $(VERILATOR_ROOT)/bin/verilator_coverage
 VERILATOR_INCLUDER = $(PYTHON3) $(VERILATOR_ROOT)/bin/verilator_includer
 VERILATOR_CCACHE_REPORT = $(PYTHON3) $(VERILATOR_ROOT)/bin/verilator_ccache_report

+######################################################################
+# CCACHE flags (via environment as no command line option available)
+
+CCACHE_SLOPPINESS ?= pch_defines,time_macros
+export CCACHE_SLOPPINESS
+
 ######################################################################
 # Make checks

@@ -67,7 +85,7 @@ VK_CPPFLAGS_ALWAYS += \
 		-DVM_TRACE_VCD=$(VM_TRACE_VCD) \
 		$(CFG_CXXFLAGS_NO_UNUSED) \

-ifeq ($(CFG_WITH_CCWARN),yes)	# Local... Else don't burden users
+ifeq ($(CFG_WITH_CCWARN),yes)  # Local... Else don't burden users
 VK_CPPFLAGS_WALL += -Wall $(CFG_CXXFLAGS_WEXTRA) -Werror
 endif

@@ -77,6 +95,8 @@ VPATH += ..
 VPATH += $(VERILATOR_ROOT)/include
 VPATH += $(VERILATOR_ROOT)/include/vltstd

+LDFLAGS  += $(CFG_LDFLAGS_VERILATED)
+
 #OPT = -ggdb -DPRINTINITSTR -DDETECTCHANGE
 #OPT = -ggdb -DPRINTINITSTR
 CPPFLAGS += $(OPT)
@@ -134,7 +154,7 @@ ifeq ($(VM_SC),1)
  LDFLAGS  += $(SYSTEMC_CXX_FLAGS) $(addprefix -L, $(SYSTEMC_LIBDIR))
  SC_LIBS   = -lsystemc
 ifneq ($(wildcard $(SYSTEMC_LIBDIR)/*numeric_bit*),)
-  # Systemc 1.2.1beta
+  # SystemC 1.2.1beta
  SC_LIBS   += -lnumeric_bit -lqt
 endif
 endif
@@ -158,11 +178,17 @@ endif
 VM_FAST += $(VM_CLASSES_FAST) $(VM_SUPPORT_FAST)
 VM_SLOW += $(VM_CLASSES_SLOW) $(VM_SUPPORT_SLOW)

+# Precompiled header filename
+VK_PCH_H = $(VM_PREFIX)__pch.h
+# Compiler read-a-precompiled-header option for precompiled header filename
+VK_PCH_I_FAST = $(CFG_CXXFLAGS_PCH_I) $(VM_PREFIX)__pch.h.fast$(CFG_GCH_IF_CLANG)
+VK_PCH_I_SLOW = $(CFG_CXXFLAGS_PCH_I) $(VM_PREFIX)__pch.h.slow$(CFG_GCH_IF_CLANG)
+
 #######################################################################
 ### Overall Objects Linking

-VK_FAST_OBJS = $(addsuffix .o, $(VM_FAST))
-VK_SLOW_OBJS = $(addsuffix .o, $(VM_SLOW))
+VK_OBJS_FAST = $(addsuffix .o, $(VM_FAST))
+VK_OBJS_SLOW = $(addsuffix .o, $(VM_SLOW))

 VK_USER_OBJS = $(addsuffix .o, $(VM_USER_CLASSES))

@@ -189,7 +215,7 @@ else
  # Parallel build: Each .cpp file by itself. This can be somewhat slower for
  # very small designs and examples, but is a lot faster for large designs.

-  VK_OBJS += $(VK_FAST_OBJS) $(VK_SLOW_OBJS)
+  VK_OBJS += $(VK_OBJS_FAST) $(VK_OBJS_SLOW)
 endif

 # When archiving just objects (.o), use single $(AR) run
@@ -233,16 +259,32 @@ $(VM_PREFIX)__ALL.a: $(VK_OBJS) $(VM_HIER_LIBS)
 ### Compile rules

 ifneq ($(VM_DEFAULT_RULES),0)
-# Anything not in $(VK_SLOW_OBJS) or $(VK_GLOBAL_OBJS), including verilated.o
-# and user files passed on the Verilator command line use this rule.
+# Compilation rule for anything not in $(VK_OBJS_FAST), $(VK_OBJS_SLOW), or
+# $(VK_GLOBAL_OBJS) including verilated.o.  This typically means user files
+# passed on the Verilator command line.
+#
+# These rules put OPT_FAST/OPT_SLOW/OPT_GLOBAL before the other flags to
+# allow USER_CPPFLAGS to override them
 %.o: %.cpp
-	$(OBJCACHE) $(CXX) $(CXXFLAGS) $(CPPFLAGS) $(OPT_FAST) -c -o $@ $<
+	$(OBJCACHE) $(CXX) $(OPT_FAST) $(CXXFLAGS) $(CPPFLAGS) -c -o $@ $<

-$(VK_SLOW_OBJS): %.o: %.cpp
-	$(OBJCACHE) $(CXX) $(CXXFLAGS) $(CPPFLAGS) $(OPT_SLOW) -c -o $@ $<
+$(VK_OBJS_FAST): %.o: %.cpp $(VK_PCH_H).fast.gch
+	$(OBJCACHE) $(CXX) $(OPT_FAST) $(CXXFLAGS) $(CPPFLAGS) $(VK_PCH_I_FAST) -c -o $@ $<
+
+$(VK_OBJS_SLOW): %.o: %.cpp $(VK_PCH_H).slow.gch
+	$(OBJCACHE) $(CXX) $(OPT_SLOW) $(CXXFLAGS) $(CPPFLAGS) $(VK_PCH_I_SLOW) -c -o $@ $<

 $(VK_GLOBAL_OBJS): %.o: %.cpp
-	$(OBJCACHE) $(CXX) $(CXXFLAGS) $(CPPFLAGS) $(OPT_GLOBAL) -c -o $@ $<
+	$(OBJCACHE) $(CXX) $(OPT_GLOBAL) $(CXXFLAGS) $(CPPFLAGS) -c -o $@ $<
+
+# Precompile a header file
+# PCH's compiler flags must match exactly the rules' above FAST/SLOW
+# arguments used for the .cpp files, or the PCH file won't be used.
+%.fast.gch: %
+	$(OBJCACHE) $(CXX) $(OPT_FAST) $(CXXFLAGS) $(CPPFLAGS) $(CFG_CXXFLAGS_PCH) $< -o $@
+%.slow.gch: %
+	$(OBJCACHE) $(CXX) $(OPT_SLOW) $(CXXFLAGS) $(CPPFLAGS) $(CFG_CXXFLAGS_PCH) $< -o $@
+
 endif

 #Default rule embedded in make:
@@ -290,19 +332,21 @@ endif

 debug-make::
 	@echo
-	@echo CXXFLAGS: $(CXXFLAGS)
 	@echo CPPFLAGS: $(CPPFLAGS)
+	@echo CXXFLAGS: $(CXXFLAGS)
 	@echo OPT_FAST: $(OPT_FAST)
 	@echo OPT_SLOW: $(OPT_SLOW)
-	@echo VM_PREFIX:  $(VM_PREFIX)
-	@echo VM_PARALLEL_BUILDS:  $(VM_PARALLEL_BUILDS)
+	@echo VK_OBJS: $(VK_OBJS)
+	@echo VK_OBJS_FAST: $(VK_OBJS_FAST)
+	@echo VK_OBJS_SLOW: $(VK_OBJS_SLOW)
 	@echo VM_CLASSES_FAST: $(VM_CLASSES_FAST)
 	@echo VM_CLASSES_SLOW: $(VM_CLASSES_SLOW)
-	@echo VM_SUPPORT_FAST: $(VM_SUPPORT_FAST)
-	@echo VM_SUPPORT_SLOW: $(VM_SUPPORT_SLOW)
 	@echo VM_GLOBAL_FAST: $(VM_GLOBAL_FAST)
 	@echo VM_GLOBAL_SLOW: $(VM_GLOBAL_SLOW)
-	@echo VK_OBJS: $(VK_OBJS)
+	@echo VM_PARALLEL_BUILDS:  $(VM_PARALLEL_BUILDS)
+	@echo VM_PREFIX:  $(VM_PREFIX)
+	@echo VM_SUPPORT_FAST: $(VM_SUPPORT_FAST)
+	@echo VM_SUPPORT_SLOW: $(VM_SUPPORT_SLOW)
 	@echo

 ######################################################################
@@ -2,7 +2,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 // can redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -208,11 +208,11 @@ private:
        // Forward to . so we have a whole word
        const std::string suffix = *bpost ? std::string{bpost + 1} : "";

-        std::string out = prefix + "*" + suffix;
+        std::string result = prefix + "*" + suffix;

-        // cout << "\nch pre="<<prefix<<"  s="<<suffix<<"\nch a="<<old<<"\nch b="<<add
-        // <<"\ncho="<<out<<endl;
-        return out;
+        // std::cout << "\nch pre=" << prefix << "  s=" << suffix << "\nch a="
+        // << old << "\nch b=" << add << "\ncho=" << result << "\n";
+        return result;
    }
    bool itemMatchesString(VerilatedCovImpItem* itemp, const std::string& match)
        VL_REQUIRES(m_mutex) {
@@ -316,7 +316,7 @@ public:
        valps[2] = page_default.c_str();

        // Keys -> strings
-        std::string keys[VerilatedCovConst::MAX_KEYS];
+        std::array<std::string, VerilatedCovConst::MAX_KEYS> keys;
        for (int i = 0; i < VerilatedCovConst::MAX_KEYS; ++i) {
            if (ckeyps[i] && ckeyps[i][0]) keys[i] = ckeyps[i];
        }
@@ -337,7 +337,7 @@ public:
            const std::string key = keys[i];
            if (!keys[i].empty()) {
                const std::string val = valps[i];
-                // cout<<"   "<<__FUNCTION__<<"  "<<key<<" = "<<val<<endl;
+                // std::cout << "   " << __FUNCTION__ << "  " << key << " = " << val << "\n";
                m_insertp->m_keys[addKeynum] = valueIndex(key);
                m_insertp->m_vals[addKeynum] = valueIndex(val);
                ++addKeynum;
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 // can redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 // can redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2009-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2009-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 // can redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -23,7 +23,6 @@

 // clang-format off

-#define __STDC_LIMIT_MACROS  // UINT64_MAX
 #include "verilated.h"
 #include "verilated_fst_c.h"

@@ -40,6 +39,7 @@
 #include <algorithm>
 #include <iterator>
 #include <sstream>
+#include <type_traits>

 #if defined(_WIN32) && !defined(__MINGW32__) && !defined(__CYGWIN__)
 # include <io.h>
@@ -50,33 +50,10 @@
 // clang-format on

 //=============================================================================
-// Check that vltscope_t matches fstScopeType
+// Check that forward declared types matches the FST API types

-static_assert(static_cast<int>(FST_ST_VCD_MODULE) == static_cast<int>(VLT_TRACE_SCOPE_MODULE),
-              "VLT_TRACE_SCOPE_MODULE mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_TASK) == static_cast<int>(VLT_TRACE_SCOPE_TASK),
-              "VLT_TRACE_SCOPE_TASK mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_FUNCTION) == static_cast<int>(VLT_TRACE_SCOPE_FUNCTION),
-              "VLT_TRACE_SCOPE_FUNCTION mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_BEGIN) == static_cast<int>(VLT_TRACE_SCOPE_BEGIN),
-              "VLT_TRACE_SCOPE_BEGIN mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_FORK) == static_cast<int>(VLT_TRACE_SCOPE_FORK),
-              "VLT_TRACE_SCOPE_FORK mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_GENERATE) == static_cast<int>(VLT_TRACE_SCOPE_GENERATE),
-              "VLT_TRACE_SCOPE_GENERATE mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_STRUCT) == static_cast<int>(VLT_TRACE_SCOPE_STRUCT),
-              "VLT_TRACE_SCOPE_STRUCT mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_UNION) == static_cast<int>(VLT_TRACE_SCOPE_UNION),
-              "VLT_TRACE_SCOPE_UNION mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_CLASS) == static_cast<int>(VLT_TRACE_SCOPE_CLASS),
-              "VLT_TRACE_SCOPE_CLASS mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_INTERFACE)
-                  == static_cast<int>(VLT_TRACE_SCOPE_INTERFACE),
-              "VLT_TRACE_SCOPE_INTERFACE mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_PACKAGE) == static_cast<int>(VLT_TRACE_SCOPE_PACKAGE),
-              "VLT_TRACE_SCOPE_PACKAGE mismatches");
-static_assert(static_cast<int>(FST_ST_VCD_PROGRAM) == static_cast<int>(VLT_TRACE_SCOPE_PROGRAM),
-              "VLT_TRACE_SCOPE_PROGRAM mismatches");
+static_assert(std::is_same<vlFstHandle, fstHandle>::value, "vlFstHandle mismatch");
+static_assert(std::is_same<vlFstEnumHandle, fstEnumHandle>::value, "vlFstHandle mismatch");

 //=============================================================================
 // Specialization of the generics for this trace format
@@ -104,19 +81,11 @@ void VerilatedFst::open(const char* filename) VL_MT_SAFE_EXCLUDES(m_mutex) {
    fstWriterSetPackType(m_fst, FST_WR_PT_LZ4);
    fstWriterSetTimescaleFromString(m_fst, timeResStr().c_str());  // lintok-begin-on-ref
    if (m_useFstWriterThread) fstWriterSetParallelMode(m_fst, 1);
+    constDump(true);  // First dump must contain the const signals
    fullDump(true);  // First dump must be full for fst

-    m_curScope.clear();
-
    Super::traceInit();

-    // Clear the scope stack
-    auto it = m_curScope.begin();
-    while (it != m_curScope.end()) {
-        fstWriterSetUpscope(m_fst);
-        it = m_curScope.erase(it);
-    }
-
    // convert m_code2symbol into an array for fast lookup
    if (!m_symbolp) {
        m_symbolp = new fstHandle[nextCode()]{0};
@@ -154,102 +123,142 @@ void VerilatedFst::declDTypeEnum(int dtypenum, const char* name, uint32_t elemen
    m_local2fstdtype[dtypenum] = enumNum;
 }

-void VerilatedFst::declare(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                           fstVarType vartype, bool array, int arraynum, bool bussed, int msb,
-                           int lsb) {
+// TODO: should return std::optional<fstScopeType>, but I can't have C++17
+static std::pair<bool, fstScopeType> toFstScopeType(VerilatedTracePrefixType type) {
+    switch (type) {
+    case VerilatedTracePrefixType::SCOPE_MODULE: return {true, FST_ST_VCD_MODULE};
+    case VerilatedTracePrefixType::SCOPE_INTERFACE: return {true, FST_ST_VCD_INTERFACE};
+    case VerilatedTracePrefixType::STRUCT_PACKED:
+    case VerilatedTracePrefixType::STRUCT_UNPACKED: return {true, FST_ST_VCD_STRUCT};
+    case VerilatedTracePrefixType::UNION_PACKED: return {true, FST_ST_VCD_UNION};
+    default: return {false, /* unused so whatever, just need a value */ FST_ST_VCD_SCOPE};
+    }
+}
+
+void VerilatedFst::pushPrefix(const std::string& name, VerilatedTracePrefixType type) {
+    const std::string newPrefix = m_prefixStack.back().first + name;
+    const auto pair = toFstScopeType(type);
+    const bool properScope = pair.first;
+    const fstScopeType scopeType = pair.second;
+    m_prefixStack.emplace_back(newPrefix + (properScope ? " " : ""), type);
+    if (properScope) {
+        const std::string scopeName = lastWord(newPrefix);
+        fstWriterSetScope(m_fst, scopeType, scopeName.c_str(), nullptr);
+    }
+}
+
+void VerilatedFst::popPrefix() {
+    const bool properScope = toFstScopeType(m_prefixStack.back().second).first;
+    if (properScope) fstWriterSetUpscope(m_fst);
+    m_prefixStack.pop_back();
+    assert(!m_prefixStack.empty());
+}
+
+void VerilatedFst::declare(uint32_t code, const char* name, int dtypenum,
+                           VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                           VerilatedTraceSigType type, bool array, int arraynum, bool bussed,
+                           int msb, int lsb) {
    const int bits = ((msb > lsb) ? (msb - lsb) : (lsb - msb)) + 1;

-    const bool enabled = Super::declCode(code, name, bits, false);
+    const std::string hierarchicalName = m_prefixStack.back().first + name;
+
+    const bool enabled = Super::declCode(code, hierarchicalName, bits);
    if (!enabled) return;

-    std::string nameasstr = namePrefix() + name;
-    std::istringstream nameiss{nameasstr};
-    std::istream_iterator<std::string> beg(nameiss);
-    std::istream_iterator<std::string> end;
-    std::list<std::string> tokens(beg, end);  // Split name
-    std::string symbol_name{tokens.back()};
-    tokens.pop_back();  // Remove symbol name from hierarchy
-    std::string tmpModName;
-
-    // Find point where current and new scope diverge
-    auto cur_it = m_curScope.begin();
-    auto new_it = tokens.begin();
-    while (cur_it != m_curScope.end() && new_it != tokens.end()) {
-        if (*cur_it != *new_it) break;
-        ++cur_it;
-        ++new_it;
-    }
-
-    // Go back to the common point
-    while (cur_it != m_curScope.end()) {
-        fstWriterSetUpscope(m_fst);
-        cur_it = m_curScope.erase(cur_it);
-    }
-
-    // Follow the hierarchy of the new variable from the common scope point
-    while (new_it != tokens.end()) {
-        if ((new_it->back() & 0x80)) {
-            tmpModName = *new_it;
-            tmpModName.pop_back();
-            // If the scope ends with a non-ASCII character, it will be 0x80 + fstScopeType
-            fstWriterSetScope(m_fst, static_cast<fstScopeType>(new_it->back() & 0x7f),
-                              tmpModName.c_str(), nullptr);
-        } else {
-            fstWriterSetScope(m_fst, FST_ST_VCD_SCOPE, new_it->c_str(), nullptr);
-        }
-        m_curScope.push_back(*new_it);
-        new_it = tokens.erase(new_it);
-    }
-
+    assert(hierarchicalName.rfind(' ') != std::string::npos);
    std::stringstream name_ss;
-    name_ss << symbol_name;
+    name_ss << lastWord(hierarchicalName);
    if (array) name_ss << "[" << arraynum << "]";
    if (bussed) name_ss << " [" << msb << ":" << lsb << "]";
-    std::string name_str = name_ss.str();
+    const std::string name_str = name_ss.str();

-    if (dtypenum > 0) {
-        const fstEnumHandle enumNum = m_local2fstdtype[dtypenum];
-        fstWriterEmitEnumTableRef(m_fst, enumNum);
+    if (dtypenum > 0) fstWriterEmitEnumTableRef(m_fst, m_local2fstdtype[dtypenum]);
+
+    fstVarDir varDir;
+    switch (direction) {
+    case VerilatedTraceSigDirection::INOUT: varDir = FST_VD_INOUT; break;
+    case VerilatedTraceSigDirection::OUTPUT: varDir = FST_VD_OUTPUT; break;
+    case VerilatedTraceSigDirection::INPUT: varDir = FST_VD_INPUT; break;
+    case VerilatedTraceSigDirection::NONE: varDir = FST_VD_IMPLICIT; break;
    }

+    fstVarType varType;
+    // Doubles have special decoding properties, so must indicate if a double
+    if (type == VerilatedTraceSigType::DOUBLE) {
+        if (kind == VerilatedTraceSigKind::PARAMETER) {
+            varType = FST_VT_VCD_REAL_PARAMETER;
+        } else {
+            varType = FST_VT_VCD_REAL;
+        }
+    }
+    // clang-format off
+    else if (kind == VerilatedTraceSigKind::PARAMETER) varType = FST_VT_VCD_PARAMETER;
+    else if (kind == VerilatedTraceSigKind::SUPPLY0) varType = FST_VT_VCD_SUPPLY0;
+    else if (kind == VerilatedTraceSigKind::SUPPLY1) varType = FST_VT_VCD_SUPPLY1;
+    else if (kind == VerilatedTraceSigKind::TRI) varType = FST_VT_VCD_TRI;
+    else if (kind == VerilatedTraceSigKind::TRI0) varType = FST_VT_VCD_TRI0;
+    else if (kind == VerilatedTraceSigKind::TRI1) varType = FST_VT_VCD_TRI1;
+    else if (kind == VerilatedTraceSigKind::WIRE) varType = FST_VT_VCD_WIRE;
+    //
+    else if (type == VerilatedTraceSigType::INTEGER) varType = FST_VT_VCD_INTEGER;
+    else if (type == VerilatedTraceSigType::BIT) varType = FST_VT_SV_BIT;
+    else if (type == VerilatedTraceSigType::LOGIC) varType = FST_VT_SV_LOGIC;
+    else if (type == VerilatedTraceSigType::INT) varType = FST_VT_SV_INT;
+    else if (type == VerilatedTraceSigType::SHORTINT) varType = FST_VT_SV_SHORTINT;
+    else if (type == VerilatedTraceSigType::LONGINT) varType = FST_VT_SV_LONGINT;
+    else if (type == VerilatedTraceSigType::BYTE) varType = FST_VT_SV_BYTE;
+    else if (type == VerilatedTraceSigType::EVENT) varType = FST_VT_VCD_EVENT;
+    else if (type == VerilatedTraceSigType::TIME) varType = FST_VT_VCD_TIME;
+    else { assert(0); /* Unreachable */ }
+    // clang-format on
+
    const auto it = vlstd::as_const(m_code2symbol).find(code);
    if (it == m_code2symbol.end()) {  // New
        m_code2symbol[code]
-            = fstWriterCreateVar(m_fst, vartype, vardir, bits, name_str.c_str(), 0);
+            = fstWriterCreateVar(m_fst, varType, varDir, bits, name_str.c_str(), 0);
    } else {  // Alias
-        fstWriterCreateVar(m_fst, vartype, vardir, bits, name_str.c_str(), it->second);
+        fstWriterCreateVar(m_fst, varType, varDir, bits, name_str.c_str(), it->second);
    }
 }

-void VerilatedFst::declEvent(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                             fstVarType vartype, bool array, int arraynum) {
-    declare(code, name, dtypenum, vardir, vartype, array, arraynum, false, 0, 0);
+void VerilatedFst::declEvent(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                             VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                             VerilatedTraceSigType type, bool array, int arraynum) {
+    declare(code, name, dtypenum, direction, kind, type, array, arraynum, false, 0, 0);
 }
-void VerilatedFst::declBit(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                           fstVarType vartype, bool array, int arraynum) {
-    declare(code, name, dtypenum, vardir, vartype, array, arraynum, false, 0, 0);
+void VerilatedFst::declBit(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                           VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                           VerilatedTraceSigType type, bool array, int arraynum) {
+    declare(code, name, dtypenum, direction, kind, type, array, arraynum, false, 0, 0);
 }
-void VerilatedFst::declBus(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                           fstVarType vartype, bool array, int arraynum, int msb, int lsb) {
-    declare(code, name, dtypenum, vardir, vartype, array, arraynum, true, msb, lsb);
+void VerilatedFst::declBus(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                           VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                           VerilatedTraceSigType type, bool array, int arraynum, int msb,
+                           int lsb) {
+    declare(code, name, dtypenum, direction, kind, type, array, arraynum, true, msb, lsb);
 }
-void VerilatedFst::declQuad(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                            fstVarType vartype, bool array, int arraynum, int msb, int lsb) {
-    declare(code, name, dtypenum, vardir, vartype, array, arraynum, true, msb, lsb);
+void VerilatedFst::declQuad(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                            VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                            VerilatedTraceSigType type, bool array, int arraynum, int msb,
+                            int lsb) {
+    declare(code, name, dtypenum, direction, kind, type, array, arraynum, true, msb, lsb);
 }
-void VerilatedFst::declArray(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                             fstVarType vartype, bool array, int arraynum, int msb, int lsb) {
-    declare(code, name, dtypenum, vardir, vartype, array, arraynum, true, msb, lsb);
+void VerilatedFst::declArray(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                             VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                             VerilatedTraceSigType type, bool array, int arraynum, int msb,
+                             int lsb) {
+    declare(code, name, dtypenum, direction, kind, type, array, arraynum, true, msb, lsb);
 }
-void VerilatedFst::declDouble(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                              fstVarType vartype, bool array, int arraynum) {
-    declare(code, name, dtypenum, vardir, vartype, array, arraynum, false, 63, 0);
+void VerilatedFst::declDouble(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                              VerilatedTraceSigDirection direction, VerilatedTraceSigKind kind,
+                              VerilatedTraceSigType type, bool array, int arraynum) {
+    declare(code, name, dtypenum, direction, kind, type, array, arraynum, false, 63, 0);
 }

 //=============================================================================
 // Get/commit trace buffer

-VerilatedFst::Buffer* VerilatedFst::getTraceBuffer() {
+VerilatedFst::Buffer* VerilatedFst::getTraceBuffer(uint32_t fidx) {
    if (offload()) return new OffloadBuffer{*this};
    return new Buffer{*this};
 }
@@ -284,7 +293,7 @@ void VerilatedFst::configure(const VerilatedTraceConfig& config) {
 // so always inline them.

 VL_ATTR_ALWINLINE
-void VerilatedFstBuffer::emitEvent(uint32_t code, VlEvent newval) {
+void VerilatedFstBuffer::emitEvent(uint32_t code, const VlEventBase* newval) {
    VL_DEBUG_IFDEF(assert(m_symbolp[code]););
    fstWriterEmitValueChange(m_fst, m_symbolp[code], "1");
 }
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 // can redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -24,13 +24,14 @@
 #include "verilated.h"
 #include "verilated_trace.h"

-#include "gtkwave/fstapi.h"
-
 #include <list>
 #include <map>
 #include <string>
 #include <vector>

+typedef uint32_t vlFstHandle;
+typedef uint32_t vlFstEnumHandle;
+
 class VerilatedFstBuffer;

 //=============================================================================
@@ -49,18 +50,22 @@ private:
    // FST specific internals

    void* m_fst = nullptr;
-    std::map<uint32_t, fstHandle> m_code2symbol;
-    std::map<int, fstEnumHandle> m_local2fstdtype;
-    std::list<std::string> m_curScope;
-    fstHandle* m_symbolp = nullptr;  // same as m_code2symbol, but as an array
+    std::map<uint32_t, vlFstHandle> m_code2symbol;
+    std::map<int, vlFstEnumHandle> m_local2fstdtype;
+    vlFstHandle* m_symbolp = nullptr;  // same as m_code2symbol, but as an array
    char* m_strbufp = nullptr;  // String buffer long enough to hold maxBits() chars

    bool m_useFstWriterThread = false;  // Whether to use the separate FST writer thread

+    // Prefixes to add to signal names/scope types
+    std::vector<std::pair<std::string, VerilatedTracePrefixType>> m_prefixStack{
+        {"", VerilatedTracePrefixType::SCOPE_MODULE}};
+
    // CONSTRUCTORS
    VL_UNCOPYABLE(VerilatedFst);
-    void declare(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                 fstVarType vartype, bool array, int arraynum, bool bussed, int msb, int lsb);
+    void declare(uint32_t code, const char* name, int dtypenum, VerilatedTraceSigDirection,
+                 VerilatedTraceSigKind, VerilatedTraceSigType, bool array, int arraynum,
+                 bool bussed, int msb, int lsb);

 protected:
    //=========================================================================
@@ -74,7 +79,7 @@ protected:
    bool preChangeDump() override { return isOpen(); }

    // Trace buffer management
-    Buffer* getTraceBuffer() override;
+    Buffer* getTraceBuffer(uint32_t fidx) override;
    void commitTraceBuffer(Buffer*) override;

    // Configure sub-class
@@ -101,18 +106,27 @@ public:
    //=========================================================================
    // Internal interface to Verilator generated code

-    void declEvent(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                   fstVarType vartype, bool array, int arraynum);
-    void declBit(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                 fstVarType vartype, bool array, int arraynum);
-    void declBus(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                 fstVarType vartype, bool array, int arraynum, int msb, int lsb);
-    void declQuad(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                  fstVarType vartype, bool array, int arraynum, int msb, int lsb);
-    void declArray(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                   fstVarType vartype, bool array, int arraynum, int msb, int lsb);
-    void declDouble(uint32_t code, const char* name, int dtypenum, fstVarDir vardir,
-                    fstVarType vartype, bool array, int arraynum);
+    void pushPrefix(const std::string&, VerilatedTracePrefixType);
+    void popPrefix();
+
+    void declEvent(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                   VerilatedTraceSigDirection, VerilatedTraceSigKind, VerilatedTraceSigType,
+                   bool array, int arraynum);
+    void declBit(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                 VerilatedTraceSigDirection, VerilatedTraceSigKind, VerilatedTraceSigType,
+                 bool array, int arraynum);
+    void declBus(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                 VerilatedTraceSigDirection, VerilatedTraceSigKind, VerilatedTraceSigType,
+                 bool array, int arraynum, int msb, int lsb);
+    void declQuad(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                  VerilatedTraceSigDirection, VerilatedTraceSigKind, VerilatedTraceSigType,
+                  bool array, int arraynum, int msb, int lsb);
+    void declArray(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                   VerilatedTraceSigDirection, VerilatedTraceSigKind, VerilatedTraceSigType,
+                   bool array, int arraynum, int msb, int lsb);
+    void declDouble(uint32_t code, uint32_t fidx, const char* name, int dtypenum,
+                    VerilatedTraceSigDirection, VerilatedTraceSigKind, VerilatedTraceSigType,
+                    bool array, int arraynum);

    void declDTypeEnum(int dtypenum, const char* name, uint32_t elements, unsigned int minValbits,
                       const char** itemNamesp, const char** itemValuesp);
@@ -149,7 +163,7 @@ class VerilatedFstBuffer VL_NOT_FINAL {
    // The FST file handle
    void* const m_fst = m_owner.m_fst;
    // code to fstHande map, as an array
-    const fstHandle* const m_symbolp = m_owner.m_symbolp;
+    const vlFstHandle* const m_symbolp = m_owner.m_symbolp;
    // String buffer long enough to hold maxBits() chars
    char* const m_strbufp = m_owner.m_strbufp;

@@ -163,7 +177,7 @@ class VerilatedFstBuffer VL_NOT_FINAL {

    // Implementations of duck-typed methods for VerilatedTraceBuffer. These are
    // called from only one place (the full* methods), so always inline them.
-    VL_ATTR_ALWINLINE void emitEvent(uint32_t code, VlEvent newval);
+    VL_ATTR_ALWINLINE void emitEvent(uint32_t code, const VlEventBase* newval);
    VL_ATTR_ALWINLINE void emitBit(uint32_t code, CData newval);
    VL_ATTR_ALWINLINE void emitCData(uint32_t code, CData newval, int bits);
    VL_ATTR_ALWINLINE void emitSData(uint32_t code, SData newval, int bits);
@@ -196,7 +210,7 @@ public:
    /// Return if file is open
    bool isOpen() const VL_MT_SAFE { return m_sptrace.isOpen(); }
    /// Open a new FST file
-    void open(const char* filename) VL_MT_SAFE { m_sptrace.open(filename); }
+    virtual void open(const char* filename) VL_MT_SAFE { m_sptrace.open(filename); }
    /// Close dump
    void close() VL_MT_SAFE { m_sptrace.close(); }
    /// Flush dump
@@ -216,12 +230,12 @@ public:

    // Set time units (s/ms, defaults to ns)
    // Users should not need to call this, as for Verilated models, these
-    // propage from the Verilated default timeunit
+    // propagate from the Verilated default timeunit
    void set_time_unit(const char* unitp) VL_MT_SAFE { m_sptrace.set_time_unit(unitp); }
    void set_time_unit(const std::string& unit) VL_MT_SAFE { m_sptrace.set_time_unit(unit); }
    // Set time resolution (s/ms, defaults to ns)
    // Users should not need to call this, as for Verilated models, these
-    // propage from the Verilated default timeprecision
+    // propagate from the Verilated default timeprecision
    void set_time_resolution(const char* unitp) VL_MT_SAFE {
        m_sptrace.set_time_resolution(unitp);
    }
@@ -3,7 +3,7 @@
 //
 // THIS MODULE IS PUBLICLY LICENSED
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you
 // can redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -1,7 +1,7 @@
 // -*- mode: C++; c-file-style: "cc-mode" -*-
 //=============================================================================
 //
-// Copyright 2001-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2001-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -12,128 +12,45 @@
 /// \file
 /// \brief Verilated tracing in FST format for SystemC header
 ///
-/// User wrapper code should use this header when creating FST SystemC
-/// traces.
+/// User wrapper code should use this header when creating FST SystemC traces.
 ///
 /// This class is not threadsafe, as the SystemC kernel is not threadsafe.
 ///
 //=============================================================================

-#ifndef VERILATOR_VERILATED_VCD_SC_H_
-#define VERILATOR_VERILATED_VCD_SC_H_
+#ifndef VERILATOR_VERILATED_FST_SC_H_
+#define VERILATOR_VERILATED_FST_SC_H_

 #include "verilatedos.h"

 #include "verilated_fst_c.h"
-#include "verilated_sc.h"
-
-#include <string>
+#include "verilated_sc_trace.h"

 //=============================================================================
 // VerilatedFstSc
-///
-/// Class representing a Verilator-friendly FST trace format registered
-/// with the SystemC simulation kernel, just like a SystemC-documented
-/// trace format.
+/// Trace file used to create FST dump for SystemC version of Verilated models. It's very similar
+/// to its C version (see the class VerilatedFstC)

-class VerilatedFstSc final : sc_trace_file, public VerilatedFstC {
+class VerilatedFstSc final : VerilatedScTraceBase, public VerilatedFstC {
    // CONSTRUCTORS
    VL_UNCOPYABLE(VerilatedFstSc);

 public:
-    /// Construct a SC trace object, and register with the SystemC kernel
    VerilatedFstSc() {
-        sc_get_curr_simcontext()->add_trace_file(this);
-        // We want to avoid a depreciated warning, but still be back compatible.
-        // Turning off the message just for this still results in an
-        // annoying "to turn off" message.
-        const sc_time t1sec{1, SC_SEC};
-        if (t1sec.to_default_time_units() != 0) {
-            const sc_time tunits{1.0 / t1sec.to_default_time_units(), SC_SEC};
-            spTrace()->set_time_unit(tunits.to_string());
-        }
-        spTrace()->set_time_resolution(sc_get_time_resolution().to_string());
-    }
-    /// Destruct, flush, and close the dump
-    ~VerilatedFstSc() override { close(); }
-
-    // METHODS - for SC kernel
-    // Called by SystemC simulate()
-    void cycle(bool delta_cycle) override {
-        if (!delta_cycle) this->dump(sc_time_stamp().to_double());
+        spTrace()->set_time_unit(VerilatedScTraceBase::getScTimeUnit());
+        spTrace()->set_time_resolution(VerilatedScTraceBase::getScTimeResolution());
    }

+    // METHODS
    // Override VerilatedFstC. Must be called after starting simulation.
-    // Note: this is not a virtual function in the base class, so no 'override'
-    virtual void open(const char* filename) VL_MT_SAFE {
-        if (VL_UNLIKELY(!sc_core::sc_get_curr_simcontext()->elaboration_done())) {
-            Verilated::scTraceBeforeElaborationError();
-        }
+    void open(const char* filename) override VL_MT_SAFE {
+        VerilatedScTraceBase::checkScElaborationDone();
        VerilatedFstC::open(filename);
    }

-private:
-    // METHODS - Fake outs for linker
-
-#ifdef NC_SYSTEMC
-    // Cadence Incisive has these as abstract functions so we must create them
-    void set_time_unit(int exponent10_seconds) override {}  // deprecated
-#endif
-    void set_time_unit(double v, sc_time_unit tu) override {}  // LCOV_EXCL_LINE
-
-    //--------------------------------------------------
-    // SystemC 2.1.v1
-
-    void write_comment(const std::string&) override {}
-    void trace(const unsigned int&, const std::string&, const char**) override {}
-
-#define DECL_TRACE_METHOD_A(tp) \
-    void trace(const tp& object, const std::string& name) override {}
-#define DECL_TRACE_METHOD_B(tp) \
-    void trace(const tp& object, const std::string& name, int width) override {}
-
-    // clang-format off
-    // Formatting matches that of sc_trace.h
-    // LCOV_EXCL_START
-#if (SYSTEMC_VERSION >= 20171012)
-    DECL_TRACE_METHOD_A( sc_event )
-    DECL_TRACE_METHOD_A( sc_time )
-#endif
-
-    DECL_TRACE_METHOD_A( bool )
-    DECL_TRACE_METHOD_A( sc_dt::sc_bit )
-    DECL_TRACE_METHOD_A( sc_dt::sc_logic )
-
-    DECL_TRACE_METHOD_B( unsigned char )
-    DECL_TRACE_METHOD_B( unsigned short )
-    DECL_TRACE_METHOD_B( unsigned int )
-    DECL_TRACE_METHOD_B( unsigned long )
-    DECL_TRACE_METHOD_B( char )
-    DECL_TRACE_METHOD_B( short )
-    DECL_TRACE_METHOD_B( int )
-    DECL_TRACE_METHOD_B( long )
-    DECL_TRACE_METHOD_B( sc_dt::int64 )
-    DECL_TRACE_METHOD_B( sc_dt::uint64 )
-
-    DECL_TRACE_METHOD_A( float )
-    DECL_TRACE_METHOD_A( double )
-    DECL_TRACE_METHOD_A( sc_dt::sc_int_base )
-    DECL_TRACE_METHOD_A( sc_dt::sc_uint_base )
-    DECL_TRACE_METHOD_A( sc_dt::sc_signed )
-    DECL_TRACE_METHOD_A( sc_dt::sc_unsigned )
-
-    DECL_TRACE_METHOD_A( sc_dt::sc_fxval )
-    DECL_TRACE_METHOD_A( sc_dt::sc_fxval_fast )
-    DECL_TRACE_METHOD_A( sc_dt::sc_fxnum )
-    DECL_TRACE_METHOD_A( sc_dt::sc_fxnum_fast )
-
-    DECL_TRACE_METHOD_A( sc_dt::sc_bv_base )
-    DECL_TRACE_METHOD_A( sc_dt::sc_lv_base )
-    // LCOV_EXCL_STOP
-    // clang-format on
-
-#undef DECL_TRACE_METHOD_A
-#undef DECL_TRACE_METHOD_B
+    // METHODS - for SC kernel
+    // Called from SystemC kernel
+    void cycle() override { VerilatedFstC::dump(sc_core::sc_time_stamp().to_double()); }
 };

 #endif  // Guard
@@ -3,7 +3,7 @@
 //
 // Code available from: https://verilator.org
 //
-// Copyright 2003-2023 by Wilson Snyder. This program is free software; you can
+// Copyright 2003-2024 by Wilson Snyder. This program is free software; you can
 // redistribute it and/or modify it under the terms of either the GNU
 // Lesser General Public License Version 3 or the Perl Artistic License
 // Version 2.0.
@@ -152,9 +152,11 @@ extern const char* vl_mc_scan_plusargs(const char* prefixp) VL_MT_SAFE;  // PLIi
 // Base macros

 // Return true if data[bit] set; not 0/1 return, but 0/non-zero return.
+// Arguments must not have side effects
 #define VL_BITISSETLIMIT_W(data, width, bit) (((bit) < (width)) && VL_BITISSET_W(data, bit))

 // Shift appropriate word by bit. Does not account for wrapping between two words
+// Argument 'bit' must not have side effects
 #define VL_BITRSHIFT_W(data, bit) ((data)[VL_BITWORD_E(bit)] >> VL_BITBIT_E(bit))

 // Create two 32-bit words from quadword
@@ -260,7 +262,7 @@ extern void _vl_debug_print_w(int lbits, WDataInP const iwp) VL_MT_SAFE;
 #if defined(SYSTEMC_VERSION)
 /// Return current simulation time
 // Already defined: extern sc_time sc_time_stamp();
-inline uint64_t vl_time_stamp64() VL_MT_SAFE { return sc_time_stamp().value(); }
+inline uint64_t vl_time_stamp64() VL_MT_SAFE { return sc_core::sc_time_stamp().value(); }
 #else  // Non-SystemC
 # if !defined(VL_TIME_CONTEXT) && !defined(VL_NO_LEGACY)
 #  ifdef VL_TIME_STAMP64
@@ -282,6 +284,8 @@ inline uint64_t vl_time_stamp64() VL_MT_SAFE {
 # endif
 #endif

+// clang-format on
+
 uint64_t VerilatedContext::time() const VL_MT_SAFE {
    // When using non-default context, fastest path is return time
    if (VL_LIKELY(m_s.m_time)) return m_s.m_time;
@@ -309,25 +313,6 @@ double vl_time_multiplier(int scale) VL_PURE;
 // Return power of 10. e.g. returns 100 if n==2
 uint64_t vl_time_pow10(int n) VL_PURE;

-#ifdef VL_DEBUG
-/// Evaluate statement if VL_DEBUG defined
-# define VL_DEBUG_IFDEF(stmt) \
-    do { \
-        stmt \
-    } while (false)
-/// Evaluate statement if VL_DEBUG defined and Verilated::debug() enabled
-# define VL_DEBUG_IF(stmt) \
-    do { \
-        if (VL_UNLIKELY(Verilated::debug())) {stmt} \
-    } while (false)
-#else
-// We intentionally do not compile the stmt to improve compile speed
-# define VL_DEBUG_IFDEF(stmt) do {} while (false)
-# define VL_DEBUG_IF(stmt) do {} while (false)
-#endif
-
-// clang-format on
-
 //=========================================================================
 // Functional macros/routines
 // These all take the form
@@ -340,10 +325,23 @@ uint64_t vl_time_pow10(int n) VL_PURE;
 //===================================================================
 // SETTING OPERATORS

+VL_ATTR_ALWINLINE
+static WDataOutP VL_MEMSET_ZERO_W(WDataOutP owp, int words) VL_MT_SAFE {
+    return static_cast<WDataOutP>(std::memset(owp, 0, words * sizeof(EData)));
+}
+VL_ATTR_ALWINLINE
+static WDataOutP VL_MEMSET_ONES_W(WDataOutP owp, int words) VL_MT_SAFE {
+    return static_cast<WDataOutP>(std::memset(owp, 0xff, words * sizeof(EData)));
+}
+VL_ATTR_ALWINLINE
+static WDataOutP VL_MEMCPY_W(WDataOutP owp, WDataInP const iwp, int words) VL_MT_SAFE {
+    return static_cast<WDataOutP>(std::memcpy(owp, iwp, words * sizeof(EData)));
+}
+
 // Output clean
 // EMIT_RULE: VL_CLEAN:  oclean=clean; obits=lbits;
-#define VL_CLEAN_II(obits, lbits, lhs) ((lhs)&VL_MASK_I(obits))
-#define VL_CLEAN_QQ(obits, lbits, lhs) ((lhs)&VL_MASK_Q(obits))
+#define VL_CLEAN_II(obits, lbits, lhs) ((lhs) & (VL_MASK_I(obits)))
+#define VL_CLEAN_QQ(obits, lbits, lhs) ((lhs) & (VL_MASK_Q(obits)))

 // EMIT_RULE: VL_ASSIGNCLEAN:  oclean=clean; obits==lbits;
 #define VL_ASSIGNCLEAN_W(obits, owp, lwp) VL_CLEAN_WW((obits), (owp), (lwp))
@@ -354,18 +352,16 @@ static inline WDataOutP _vl_clean_inplace_w(int obits, WDataOutP owp) VL_MT_SAFE
 }
 static inline WDataOutP VL_CLEAN_WW(int obits, WDataOutP owp, WDataInP const lwp) VL_MT_SAFE {
    const int words = VL_WORDS_I(obits);
-    for (int i = 0; (i < (words - 1)); ++i) owp[i] = lwp[i];
+    VL_MEMCPY_W(owp, lwp, words - 1);
    owp[words - 1] = lwp[words - 1] & VL_MASK_E(obits);
    return owp;
 }
 static inline WDataOutP VL_ZERO_W(int obits, WDataOutP owp) VL_MT_SAFE {
-    const int words = VL_WORDS_I(obits);
-    for (int i = 0; i < words; ++i) owp[i] = 0;
-    return owp;
+    return VL_MEMSET_ZERO_W(owp, VL_WORDS_I(obits));
 }
 static inline WDataOutP VL_ALLONES_W(int obits, WDataOutP owp) VL_MT_SAFE {
    const int words = VL_WORDS_I(obits);
-    for (int i = 0; i < (words - 1); ++i) owp[i] = ~VL_EUL(0);
+    VL_MEMSET_ONES_W(owp, words - 1);
    owp[words - 1] = VL_MASK_E(obits);
    return owp;
 }
@@ -374,9 +370,7 @@ static inline WDataOutP VL_ALLONES_W(int obits, WDataOutP owp) VL_MT_SAFE {
 // For now, we always have a clean rhs.
 // Note: If a ASSIGN isn't clean, use VL_ASSIGNCLEAN instead to do the same thing.
 static inline WDataOutP VL_ASSIGN_W(int obits, WDataOutP owp, WDataInP const lwp) VL_MT_SAFE {
-    const int words = VL_WORDS_I(obits);
-    for (int i = 0; i < words; ++i) owp[i] = lwp[i];
-    return owp;
+    return VL_MEMCPY_W(owp, lwp, VL_WORDS_I(obits));
 }

 // EMIT_RULE: VL_ASSIGNBIT:  rclean=clean;
@@ -447,7 +441,7 @@ static inline void VL_ASSIGNBIT_WO(int bit, WDataOutP owp) VL_MT_SAFE {
 #define VL_ASSIGN_WSB(obits, owp, svar) \
    { \
        const int words = VL_WORDS_I(obits); \
-        sc_biguint<(obits)> _butemp = (svar).read(); \
+        sc_dt::sc_biguint<(obits)> _butemp = (svar).read(); \
        uint32_t* chunkp = _butemp.get_raw(); \
        int32_t lsb = 0; \
        while (lsb < obits - BITS_PER_DIGIT) { \
@@ -473,20 +467,20 @@ static inline void VL_ASSIGNBIT_WO(int bit, WDataOutP owp) VL_MT_SAFE {

 #define VL_ASSIGN_SWI(obits, svar, rd) \
    { \
-        sc_bv<(obits)> _bvtemp; \
+        sc_dt::sc_bv<(obits)> _bvtemp; \
        _bvtemp.set_word(0, (rd)); \
        (svar).write(_bvtemp); \
    }
 #define VL_ASSIGN_SWQ(obits, svar, rd) \
    { \
-        sc_bv<(obits)> _bvtemp; \
+        sc_dt::sc_bv<(obits)> _bvtemp; \
        _bvtemp.set_word(0, static_cast<IData>(rd)); \
        _bvtemp.set_word(1, static_cast<IData>((rd) >> VL_IDATASIZE)); \
        (svar).write(_bvtemp); \
    }
 #define VL_ASSIGN_SWW(obits, svar, rwp) \
    { \
-        sc_bv<(obits)> _bvtemp; \
+        sc_dt::sc_bv<(obits)> _bvtemp; \
        for (int i = 0; i < VL_WORDS_I(obits); ++i) _bvtemp.set_word(i, (rwp)[i]); \
        (svar).write(_bvtemp); \
    }
@@ -502,7 +496,7 @@ static inline void VL_ASSIGNBIT_WO(int bit, WDataOutP owp) VL_MT_SAFE {
 #define VL_SC_BITS_PER_DIGIT 30  // This comes from sc_nbdefs.h BITS_PER_DIGIT
 #define VL_ASSIGN_SBW(obits, svar, rwp) \
    { \
-        sc_biguint<(obits)> _butemp; \
+        sc_dt::sc_biguint<(obits)> _butemp; \
        int32_t lsb = 0; \
        uint32_t* chunkp = _butemp.get_raw(); \
        while (lsb + VL_SC_BITS_PER_DIGIT < (obits)) { \
@@ -534,19 +528,20 @@ static inline void VL_ASSIGNBIT_WO(int bit, WDataOutP owp) VL_MT_SAFE {
 static inline WDataOutP VL_EXTEND_WI(int obits, int, WDataOutP owp, IData ld) VL_MT_SAFE {
    // Note for extracts that obits != lbits
    owp[0] = ld;
-    for (int i = 1; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + 1, VL_WORDS_I(obits) - 1);
    return owp;
 }
 static inline WDataOutP VL_EXTEND_WQ(int obits, int, WDataOutP owp, QData ld) VL_MT_SAFE {
    VL_SET_WQ(owp, ld);
-    for (int i = VL_WQ_WORDS_E; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + VL_WQ_WORDS_E, VL_WORDS_I(obits) - VL_WQ_WORDS_E);
    return owp;
 }
 static inline WDataOutP VL_EXTEND_WW(int obits, int lbits, WDataOutP owp,
                                     WDataInP const lwp) VL_MT_SAFE {
-    for (int i = 0; i < VL_WORDS_I(lbits); ++i) owp[i] = lwp[i];
-    for (int i = VL_WORDS_I(lbits); i < VL_WORDS_I(obits); ++i) owp[i] = 0;
-    return owp;
+    const int lwords = VL_WORDS_I(lbits);
+    VL_PREFETCH_RD(lwp);
+    VL_MEMSET_ZERO_W(owp + lwords, VL_WORDS_I(obits) - lwords);
+    return VL_MEMCPY_W(owp, lwp, lwords);
 }

 // EMIT_RULE: VL_EXTENDS:  oclean=*dirty*; obits=lbits;
@@ -562,26 +557,37 @@ static inline QData VL_EXTENDS_QQ(int, int lbits, QData lhs) VL_PURE {
 }

 static inline WDataOutP VL_EXTENDS_WI(int obits, int lbits, WDataOutP owp, IData ld) VL_MT_SAFE {
-    const EData sign = VL_SIGNONES_E(lbits, static_cast<EData>(ld));
-    owp[0] = ld | (sign & ~VL_MASK_E(lbits));
-    for (int i = 1; i < VL_WORDS_I(obits); ++i) owp[i] = sign;
+    owp[0] = ld;
+    if (VL_SIGN_E(lbits, owp[0])) {
+        owp[0] |= ~VL_MASK_E(lbits);
+        VL_MEMSET_ONES_W(owp + 1, VL_WORDS_I(obits) - 1);
+    } else {
+        VL_MEMSET_ZERO_W(owp + 1, VL_WORDS_I(obits) - 1);
+    }
    return owp;
 }
 static inline WDataOutP VL_EXTENDS_WQ(int obits, int lbits, WDataOutP owp, QData ld) VL_MT_SAFE {
    VL_SET_WQ(owp, ld);
-    const EData sign = VL_SIGNONES_E(lbits, owp[1]);
-    owp[1] |= sign & ~VL_MASK_E(lbits);
-    for (int i = VL_WQ_WORDS_E; i < VL_WORDS_I(obits); ++i) owp[i] = sign;
+    if (VL_SIGN_E(lbits, owp[1])) {
+        owp[1] |= ~VL_MASK_E(lbits);
+        VL_MEMSET_ONES_W(owp + VL_WQ_WORDS_E, VL_WORDS_I(obits) - VL_WQ_WORDS_E);
+    } else {
+        VL_MEMSET_ZERO_W(owp + VL_WQ_WORDS_E, VL_WORDS_I(obits) - VL_WQ_WORDS_E);
+    }
    return owp;
 }
 static inline WDataOutP VL_EXTENDS_WW(int obits, int lbits, WDataOutP owp,
                                      WDataInP const lwp) VL_MT_SAFE {
-    for (int i = 0; i < VL_WORDS_I(lbits) - 1; ++i) owp[i] = lwp[i];
-    const int lmsw = VL_WORDS_I(lbits) - 1;
-    const EData sign = VL_SIGNONES_E(lbits, lwp[lmsw]);
-    owp[lmsw] = lwp[lmsw] | (sign & ~VL_MASK_E(lbits));
-    for (int i = VL_WORDS_I(lbits); i < VL_WORDS_I(obits); ++i) owp[i] = sign;
-    return owp;
+    const int lwords = VL_WORDS_I(lbits);
+    VL_PREFETCH_RD(lwp);
+    owp[lwords - 1] = lwp[lwords - 1];
+    if (VL_SIGN_E(lbits, lwp[lwords - 1])) {
+        owp[lwords - 1] |= ~VL_MASK_E(lbits);
+        VL_MEMSET_ONES_W(owp + lwords, VL_WORDS_I(obits) - lwords);
+    } else {
+        VL_MEMSET_ZERO_W(owp + lwords, VL_WORDS_I(obits) - lwords);
+    }
+    return VL_MEMCPY_W(owp, lwp, lwords - 1);
 }

 //===================================================================
@@ -758,6 +764,7 @@ static inline IData VL_ONEHOT0_W(int words, WDataInP const lwp) VL_PURE {

 static inline IData VL_CLOG2_I(IData lhs) VL_PURE {
    // There are faster algorithms, or fls GCC4 builtins, but rarely used
+    // In C++20 there will be std::bit_width(lhs) - 1
    if (VL_UNLIKELY(!lhs)) return 0;
    --lhs;
    int shifts = 0;
@@ -953,11 +960,19 @@ static inline void VL_NEGATE_INPLACE_W(int words, WDataOutP owp_lwp) VL_MT_SAFE
 // EMIT_RULE: VL_MUL:    oclean=dirty; lclean==clean; rclean==clean;
 // EMIT_RULE: VL_DIV:    oclean=dirty; lclean==clean; rclean==clean;
 // EMIT_RULE: VL_MODDIV: oclean=dirty; lclean==clean; rclean==clean;
-#define VL_DIV_III(lbits, lhs, rhs) (((rhs) == 0) ? 0 : (lhs) / (rhs))
-#define VL_DIV_QQQ(lbits, lhs, rhs) (((rhs) == 0) ? 0 : (lhs) / (rhs))
+static inline IData VL_DIV_III(int lbits, IData lhs, IData rhs) {
+    return (rhs == 0) ? 0 : lhs / rhs;
+}
+static inline QData VL_DIV_QQQ(int lbits, QData lhs, QData rhs) {
+    return (rhs == 0) ? 0 : lhs / rhs;
+}
 #define VL_DIV_WWW(lbits, owp, lwp, rwp) (_vl_moddiv_w(lbits, owp, lwp, rwp, 0))
-#define VL_MODDIV_III(lbits, lhs, rhs) (((rhs) == 0) ? 0 : (lhs) % (rhs))
-#define VL_MODDIV_QQQ(lbits, lhs, rhs) (((rhs) == 0) ? 0 : (lhs) % (rhs))
+static inline IData VL_MODDIV_III(int lbits, IData lhs, IData rhs) {
+    return (rhs == 0) ? 0 : lhs % rhs;
+}
+static inline QData VL_MODDIV_QQQ(int lbits, QData lhs, QData rhs) {
+    return (rhs == 0) ? 0 : lhs % rhs;
+}
 #define VL_MODDIV_WWW(lbits, owp, lwp, rwp) (_vl_moddiv_w(lbits, owp, lwp, rwp, 1))

 static inline WDataOutP VL_ADD_W(int words, WDataOutP owp, WDataInP const lwp,
@@ -1017,6 +1032,7 @@ static inline QData VL_MULS_QQQ(int lbits, QData lhs, QData rhs) VL_PURE {
 static inline WDataOutP VL_MULS_WWW(int lbits, WDataOutP owp, WDataInP const lwp,
                                    WDataInP const rwp) VL_MT_SAFE {
    const int words = VL_WORDS_I(lbits);
+    VL_DEBUG_IFDEF(assert(words <= VL_MULS_MAX_WORDS););
    // cppcheck-suppress variableScope
    WData lwstore[VL_MULS_MAX_WORDS];  // Fixed size, as MSVC++ doesn't allow [words] here
    // cppcheck-suppress variableScope
@@ -1087,21 +1103,22 @@ static inline QData VL_MODDIVS_QQQ(int lbits, QData lhs, QData rhs) VL_PURE {

 static inline WDataOutP VL_DIVS_WWW(int lbits, WDataOutP owp, WDataInP const lwp,
                                    WDataInP const rwp) VL_MT_SAFE {
-    const int words = VL_WORDS_I(lbits);
-    const EData lsign = VL_SIGN_E(lbits, lwp[words - 1]);
-    const EData rsign = VL_SIGN_E(lbits, rwp[words - 1]);
+    const int lwords = VL_WORDS_I(lbits);
+    const EData lsign = VL_SIGN_E(lbits, lwp[lwords - 1]);
+    const EData rsign = VL_SIGN_E(lbits, rwp[lwords - 1]);
+    VL_DEBUG_IFDEF(assert(lwords <= VL_MULS_MAX_WORDS););
    // cppcheck-suppress variableScope
    WData lwstore[VL_MULS_MAX_WORDS];  // Fixed size, as MSVC++ doesn't allow [words] here
    // cppcheck-suppress variableScope
    WData rwstore[VL_MULS_MAX_WORDS];
    WDataInP ltup = lwp;
    WDataInP rtup = rwp;
-    if (lsign) ltup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(VL_WORDS_I(lbits), lwstore, lwp));
-    if (rsign) rtup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(VL_WORDS_I(lbits), rwstore, rwp));
+    if (lsign) ltup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(lwords, lwstore, lwp));
+    if (rsign) rtup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(lwords, rwstore, rwp));
    if ((lsign && !rsign) || (!lsign && rsign)) {
        WData qNoSign[VL_MULS_MAX_WORDS];
        VL_DIV_WWW(lbits, qNoSign, ltup, rtup);
-        _vl_clean_inplace_w(lbits, VL_NEGATE_W(VL_WORDS_I(lbits), owp, qNoSign));
+        _vl_clean_inplace_w(lbits, VL_NEGATE_W(lwords, owp, qNoSign));
        return owp;
    } else {
        return VL_DIV_WWW(lbits, owp, ltup, rtup);
@@ -1109,21 +1126,22 @@ static inline WDataOutP VL_DIVS_WWW(int lbits, WDataOutP owp, WDataInP const lwp
 }
 static inline WDataOutP VL_MODDIVS_WWW(int lbits, WDataOutP owp, WDataInP const lwp,
                                       WDataInP const rwp) VL_MT_SAFE {
-    const int words = VL_WORDS_I(lbits);
-    const EData lsign = VL_SIGN_E(lbits, lwp[words - 1]);
-    const EData rsign = VL_SIGN_E(lbits, rwp[words - 1]);
+    const int lwords = VL_WORDS_I(lbits);
+    const EData lsign = VL_SIGN_E(lbits, lwp[lwords - 1]);
+    const EData rsign = VL_SIGN_E(lbits, rwp[lwords - 1]);
+    VL_DEBUG_IFDEF(assert(lwords <= VL_MULS_MAX_WORDS););
    // cppcheck-suppress variableScope
    WData lwstore[VL_MULS_MAX_WORDS];  // Fixed size, as MSVC++ doesn't allow [words] here
    // cppcheck-suppress variableScope
    WData rwstore[VL_MULS_MAX_WORDS];
    WDataInP ltup = lwp;
    WDataInP rtup = rwp;
-    if (lsign) ltup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(VL_WORDS_I(lbits), lwstore, lwp));
-    if (rsign) rtup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(VL_WORDS_I(lbits), rwstore, rwp));
+    if (lsign) ltup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(lwords, lwstore, lwp));
+    if (rsign) rtup = _vl_clean_inplace_w(lbits, VL_NEGATE_W(lwords, rwstore, rwp));
    if (lsign) {  // Only dividend sign matters for modulus
        WData qNoSign[VL_MULS_MAX_WORDS];
        VL_MODDIV_WWW(lbits, qNoSign, ltup, rtup);
-        _vl_clean_inplace_w(lbits, VL_NEGATE_W(VL_WORDS_I(lbits), owp, qNoSign));
+        _vl_clean_inplace_w(lbits, VL_NEGATE_W(lwords, owp, qNoSign));
        return owp;
    } else {
        return VL_MODDIV_WWW(lbits, owp, ltup, rtup);
@@ -1426,7 +1444,7 @@ static inline IData VL_STREAML_FAST_III(int lbits, IData ld, IData rd_log2) VL_P
    if (rd_log2) {
        const uint32_t lbitsFloor = lbits & ~VL_MASK_I(rd_log2);  // max multiple of rd <= lbits
        const uint32_t lbitsRem = lbits - lbitsFloor;  // number of bits in most-sig slice (MSS)
-        const IData msbMask = VL_MASK_I(lbitsRem) << lbitsFloor;  // mask to sel only bits in MSS
+        const IData msbMask = lbitsFloor == 32 ? 0UL : VL_MASK_I(lbitsRem) << lbitsFloor;
        ret = (ret & ~msbMask) | ((ret & msbMask) << ((VL_UL(1) << rd_log2) - lbitsRem));
    }
    switch (rd_log2) {
@@ -1515,6 +1533,28 @@ static inline WDataOutP VL_STREAML_WWI(int lbits, WDataOutP owp, WDataInP const
    return owp;
 }

+template <typename T>
+static inline void VL_ASSIGN_DYN_Q(VlQueue<T>& q, int elem_size, int lbits, QData from) {
+    const int size = (lbits + elem_size - 1) / elem_size;
+    q.renew(size);
+    const QData mask = VL_MASK_Q(elem_size);
+    for (int i = 0; i < size; ++i) q.at(i) = (T)((from >> (i * elem_size)) & mask);
+}
+
+template <typename T>
+static inline IData VL_DYN_TO_I(const VlQueue<T>& q, int elem_size) {
+    IData ret = 0;
+    for (int i = 0; i < q.size(); ++i) ret |= q.at(i) << (i * elem_size);
+    return ret;
+}
+
+template <typename T>
+static inline QData VL_DYN_TO_Q(const VlQueue<T>& q, int elem_size) {
+    QData ret = 0;
+    for (int i = 0; i < q.size(); ++i) ret |= q.at(i) << (i * elem_size);
+    return ret;
+}
+
 // Because concats are common and wide, it's valuable to always have a clean output.
 // Thus we specify inputs must be clean, so we don't need to clean the output.
 // Note the bit shifts are always constants, so the adds in these constify out.
@@ -1533,63 +1573,66 @@ static inline WDataOutP VL_STREAML_WWI(int lbits, WDataOutP owp, WDataInP const
 static inline WDataOutP VL_CONCAT_WII(int obits, int lbits, int rbits, WDataOutP owp, IData ld,
                                      IData rd) VL_MT_SAFE {
    owp[0] = rd;
-    for (int i = 1; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + 1, VL_WORDS_I(obits) - 1);
    _vl_insert_WI(owp, ld, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WWI(int obits, int lbits, int rbits, WDataOutP owp,
                                      WDataInP const lwp, IData rd) VL_MT_SAFE {
    owp[0] = rd;
-    for (int i = 1; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + 1, VL_WORDS_I(obits) - 1);
    _vl_insert_WW(owp, lwp, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WIW(int obits, int lbits, int rbits, WDataOutP owp, IData ld,
                                      WDataInP const rwp) VL_MT_SAFE {
-    for (int i = 0; i < VL_WORDS_I(rbits); ++i) owp[i] = rwp[i];
-    for (int i = VL_WORDS_I(rbits); i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    const int rwords = VL_WORDS_I(rbits);
+    VL_MEMCPY_W(owp, rwp, rwords);
+    VL_MEMSET_ZERO_W(owp + rwords, VL_WORDS_I(obits) - rwords);
    _vl_insert_WI(owp, ld, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WIQ(int obits, int lbits, int rbits, WDataOutP owp, IData ld,
                                      QData rd) VL_MT_SAFE {
    VL_SET_WQ(owp, rd);
-    for (int i = VL_WQ_WORDS_E; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + VL_WQ_WORDS_E, VL_WORDS_I(obits) - VL_WQ_WORDS_E);
    _vl_insert_WI(owp, ld, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WQI(int obits, int lbits, int rbits, WDataOutP owp, QData ld,
                                      IData rd) VL_MT_SAFE {
    owp[0] = rd;
-    for (int i = 1; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + 1, VL_WORDS_I(obits) - 1);
    _vl_insert_WQ(owp, ld, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WQQ(int obits, int lbits, int rbits, WDataOutP owp, QData ld,
                                      QData rd) VL_MT_SAFE {
    VL_SET_WQ(owp, rd);
-    for (int i = VL_WQ_WORDS_E; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + VL_WQ_WORDS_E, VL_WORDS_I(obits) - VL_WQ_WORDS_E);
    _vl_insert_WQ(owp, ld, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WWQ(int obits, int lbits, int rbits, WDataOutP owp,
                                      WDataInP const lwp, QData rd) VL_MT_SAFE {
    VL_SET_WQ(owp, rd);
-    for (int i = VL_WQ_WORDS_E; i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    VL_MEMSET_ZERO_W(owp + VL_WQ_WORDS_E, VL_WORDS_I(obits) - VL_WQ_WORDS_E);
    _vl_insert_WW(owp, lwp, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WQW(int obits, int lbits, int rbits, WDataOutP owp, QData ld,
                                      WDataInP const rwp) VL_MT_SAFE {
-    for (int i = 0; i < VL_WORDS_I(rbits); ++i) owp[i] = rwp[i];
-    for (int i = VL_WORDS_I(rbits); i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    const int rwords = VL_WORDS_I(rbits);
+    VL_MEMCPY_W(owp, rwp, rwords);
+    VL_MEMSET_ZERO_W(owp + rwords, VL_WORDS_I(obits) - rwords);
    _vl_insert_WQ(owp, ld, rbits + lbits - 1, rbits);
    return owp;
 }
 static inline WDataOutP VL_CONCAT_WWW(int obits, int lbits, int rbits, WDataOutP owp,
                                      WDataInP const lwp, WDataInP const rwp) VL_MT_SAFE {
-    for (int i = 0; i < VL_WORDS_I(rbits); ++i) owp[i] = rwp[i];
-    for (int i = VL_WORDS_I(rbits); i < VL_WORDS_I(obits); ++i) owp[i] = 0;
+    const int rwords = VL_WORDS_I(rbits);
+    VL_MEMCPY_W(owp, rwp, rwords);
+    VL_MEMSET_ZERO_W(owp + rwords, VL_WORDS_I(obits) - rwords);
    _vl_insert_WW(owp, lwp, rbits + lbits - 1, rbits);
    return owp;
 }
@@ -1614,6 +1657,24 @@ static inline void _vl_shiftl_inplace_w(int obits, WDataOutP iowp,
 // EMIT_RULE: VL_SHIFTL:  oclean=lclean; rclean==clean;
 // Important: Unlike most other funcs, the shift might well be a computed
 // expression.  Thus consider this when optimizing.  (And perhaps have 2 funcs?)
+// If RHS (rd/rwp) is larger than the output, zeros (or all ones for >>>) must be returned
+// (This corresponds to AstShift*Ovr Ast nodes)
+static inline IData VL_SHIFTL_III(int obits, int, int, IData lhs, IData rhs) VL_MT_SAFE {
+    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return 0;
+    return lhs << rhs;  // Small is common so not clean return
+}
+static inline IData VL_SHIFTL_IIQ(int obits, int, int, IData lhs, QData rhs) VL_MT_SAFE {
+    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return 0;
+    return VL_CLEAN_II(obits, obits, lhs << rhs);
+}
+static inline QData VL_SHIFTL_QQI(int obits, int, int, QData lhs, IData rhs) VL_MT_SAFE {
+    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return 0;
+    return lhs << rhs;  // Small is common so not clean return
+}
+static inline QData VL_SHIFTL_QQQ(int obits, int, int, QData lhs, QData rhs) VL_MT_SAFE {
+    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return 0;
+    return VL_CLEAN_QQ(obits, obits, lhs << rhs);
+}
 static inline WDataOutP VL_SHIFTL_WWI(int obits, int, int, WDataOutP owp, WDataInP const lwp,
                                      IData rd) VL_MT_SAFE {
    const int word_shift = VL_BITWORD_E(rd);
@@ -1651,11 +1712,7 @@ static inline IData VL_SHIFTL_IIW(int obits, int, int rbits, IData lhs,
            return 0;
        }
    }
-    return VL_CLEAN_II(obits, obits, lhs << rwp[0]);
-}
-static inline IData VL_SHIFTL_IIQ(int obits, int, int, IData lhs, QData rhs) VL_MT_SAFE {
-    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return 0;
-    return VL_CLEAN_II(obits, obits, lhs << rhs);
+    return VL_SHIFTL_III(obits, obits, 32, lhs, rwp[0]);
 }
 static inline QData VL_SHIFTL_QQW(int obits, int, int rbits, QData lhs,
                                  WDataInP const rwp) VL_MT_SAFE {
@@ -1665,16 +1722,28 @@ static inline QData VL_SHIFTL_QQW(int obits, int, int rbits, QData lhs,
        }
    }
    // Above checks rwp[1]==0 so not needed in below shift
-    return VL_CLEAN_QQ(obits, obits, lhs << (static_cast<QData>(rwp[0])));
-}
-static inline QData VL_SHIFTL_QQQ(int obits, int, int, QData lhs, QData rhs) VL_MT_SAFE {
-    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return 0;
-    return VL_CLEAN_QQ(obits, obits, lhs << rhs);
+    return VL_SHIFTL_QQI(obits, obits, 32, lhs, rwp[0]);
 }

 // EMIT_RULE: VL_SHIFTR:  oclean=lclean; rclean==clean;
 // Important: Unlike most other funcs, the shift might well be a computed
 // expression.  Thus consider this when optimizing.  (And perhaps have 2 funcs?)
+static inline IData VL_SHIFTR_III(int obits, int, int, IData lhs, IData rhs) VL_PURE {
+    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return 0;
+    return lhs >> rhs;  // Small is common so assumed not clean
+}
+static inline IData VL_SHIFTR_IIQ(int obits, int, int, IData lhs, QData rhs) VL_PURE {
+    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return 0;
+    return VL_CLEAN_QQ(obits, obits, lhs >> rhs);
+}
+static inline QData VL_SHIFTR_QQI(int obits, int, int, QData lhs, IData rhs) VL_PURE {
+    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return 0;
+    return lhs >> rhs;  // Small is common so assumed not clean
+}
+static inline QData VL_SHIFTR_QQQ(int obits, int, int, QData lhs, QData rhs) VL_PURE {
+    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return 0;
+    return VL_CLEAN_QQ(obits, obits, lhs >> rhs);
+}
 static inline WDataOutP VL_SHIFTR_WWI(int obits, int, int, WDataOutP owp, WDataInP const lwp,
                                      IData rd) VL_MT_SAFE {
    const int word_shift = VL_BITWORD_E(rd);  // Maybe 0
@@ -1718,29 +1787,16 @@ static inline WDataOutP VL_SHIFTR_WWQ(int obits, int lbits, int rbits, WDataOutP
 static inline IData VL_SHIFTR_IIW(int obits, int, int rbits, IData lhs,
                                  WDataInP const rwp) VL_PURE {
    for (int i = 1; i < VL_WORDS_I(rbits); ++i) {
-        if (VL_UNLIKELY(rwp[i])) {  // Huge shift 1>>32 or more
-            return 0;
-        }
+        if (VL_UNLIKELY(rwp[i])) return 0;  // Huge shift 1>>32 or more
    }
-    return VL_CLEAN_II(obits, obits, lhs >> rwp[0]);
+    return VL_SHIFTR_III(obits, obits, 32, lhs, rwp[0]);
 }
 static inline QData VL_SHIFTR_QQW(int obits, int, int rbits, QData lhs,
                                  WDataInP const rwp) VL_PURE {
    for (int i = 1; i < VL_WORDS_I(rbits); ++i) {
-        if (VL_UNLIKELY(rwp[i])) {  // Huge shift 1>>32 or more
-            return 0;
-        }
+        if (VL_UNLIKELY(rwp[i])) return 0;  // Huge shift 1>>32 or more
    }
-    // Above checks rwp[1]==0 so not needed in below shift
-    return VL_CLEAN_QQ(obits, obits, lhs >> (static_cast<QData>(rwp[0])));
-}
-static inline IData VL_SHIFTR_IIQ(int obits, int, int, IData lhs, QData rhs) VL_PURE {
-    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return 0;
-    return VL_CLEAN_QQ(obits, obits, lhs >> rhs);
-}
-static inline QData VL_SHIFTR_QQQ(int obits, int, int, QData lhs, QData rhs) VL_PURE {
-    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return 0;
-    return VL_CLEAN_QQ(obits, obits, lhs >> rhs);
+    return VL_SHIFTR_QQI(obits, obits, 32, lhs, rwp[0]);
 }

 // EMIT_RULE: VL_SHIFTRS:  oclean=false; lclean=clean, rclean==clean;
@@ -1750,11 +1806,13 @@ static inline IData VL_SHIFTRS_III(int obits, int lbits, int, IData lhs, IData r
    // must use lbits for sign; lbits might != obits,
    // an EXTEND(SHIFTRS(...)) can became a SHIFTRS(...) within same 32/64 bit word length
    const IData sign = -(lhs >> (lbits - 1));  // ffff_ffff if negative
+    if (VL_UNLIKELY(rhs >= VL_IDATASIZE)) return sign & VL_MASK_I(obits);
    const IData signext = ~(VL_MASK_I(lbits) >> rhs);  // One with bits where we've shifted "past"
    return (lhs >> rhs) | (sign & VL_CLEAN_II(obits, obits, signext));
 }
 static inline QData VL_SHIFTRS_QQI(int obits, int lbits, int, QData lhs, IData rhs) VL_PURE {
    const QData sign = -(lhs >> (lbits - 1));
+    if (VL_UNLIKELY(rhs >= VL_QUADSIZE)) return sign & VL_MASK_Q(obits);
    const QData signext = ~(VL_MASK_Q(lbits) >> rhs);
    return (lhs >> rhs) | (sign & VL_CLEAN_QQ(obits, obits, signext));
 }
@@ -1798,10 +1856,13 @@ static inline WDataOutP VL_SHIFTRS_WWW(int obits, int lbits, int rbits, WDataOut
    EData overshift = 0;  // Huge shift 1>>32 or more
    for (int i = 1; i < VL_WORDS_I(rbits); ++i) overshift |= rwp[i];
    if (VL_UNLIKELY(overshift || rwp[0] >= static_cast<IData>(obits))) {
-        const int lmsw = VL_WORDS_I(obits) - 1;
-        const EData sign = VL_SIGNONES_E(lbits, lwp[lmsw]);
-        for (int j = 0; j <= lmsw; ++j) owp[j] = sign;
-        owp[lmsw] &= VL_MASK_E(lbits);
+        const int owords = VL_WORDS_I(obits);
+        if (VL_SIGN_E(lbits, lwp[owords - 1])) {
+            VL_MEMSET_ONES_W(owp, owords);
+            owp[owords - 1] &= VL_MASK_E(lbits);
+        } else {
+            VL_MEMSET_ZERO_W(owp, owords);
+        }
        return owp;
    }
    return VL_SHIFTRS_WWI(obits, lbits, 32, owp, lwp, rwp[0]);
@@ -2012,9 +2073,7 @@ static inline void VL_ASSIGNSEL_WW(int rbits, int obits, int lsb, WDataOutP iowp

 static inline WDataOutP VL_COND_WIWW(int obits, WDataOutP owp, int cond, WDataInP const w1p,
                                     WDataInP const w2p) VL_MT_SAFE {
-    const int words = VL_WORDS_I(obits);
-    for (int i = 0; i < words; ++i) owp[i] = cond ? w1p[i] : w2p[i];
-    return owp;
+    return VL_MEMCPY_W(owp, cond ? w1p : w2p, VL_WORDS_I(obits));
 }

 //======================================================================
@@ -2027,7 +2086,7 @@ static inline WDataOutP VL_COND_WIWW(int obits, WDataOutP owp, int cond, WDataIn
 // If changing the number of functions here, also change EMITCINLINES_NUM_CONSTW

 #define VL_C_END_(obits, wordsSet) \
-    for (int i = (wordsSet); i < VL_WORDS_I(obits); ++i) o[i] = 0; \
+    VL_MEMSET_ZERO_W(o + (wordsSet), VL_WORDS_I(obits) - (wordsSet)); \
    return o

 // clang-format off
@@ -2078,60 +2137,60 @@ static inline WDataOutP VL_CONST_W_8X(int obits, WDataOutP o,
    VL_C_END_(obits, 8);
 }
 //
-static inline WDataOutP VL_CONSTHI_W_1X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_1X(int obits, int lsb, WDataOutP o,
                                        EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 1);
 }
-static inline WDataOutP VL_CONSTHI_W_2X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_2X(int obits, int lsb, WDataOutP o,
                                        EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 2);
 }
-static inline WDataOutP VL_CONSTHI_W_3X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_3X(int obits, int lsb, WDataOutP o,
                                        EData d2, EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;  o[2] = d2;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;  ohi[2] = d2;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 3);
 }
-static inline WDataOutP VL_CONSTHI_W_4X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_4X(int obits, int lsb, WDataOutP o,
                                        EData d3, EData d2, EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;  o[2] = d2;  o[3] = d3;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;  ohi[2] = d2;  ohi[3] = d3;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 4);
 }
-static inline WDataOutP VL_CONSTHI_W_5X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_5X(int obits, int lsb, WDataOutP o,
                                        EData d4,
                                        EData d3, EData d2, EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;  o[2] = d2;  o[3] = d3;
-    o[4] = d4;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;  ohi[2] = d2;  ohi[3] = d3;
+    ohi[4] = d4;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 5);
 }
-static inline WDataOutP VL_CONSTHI_W_6X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_6X(int obits, int lsb, WDataOutP o,
                                        EData d5, EData d4,
                                        EData d3, EData d2, EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;  o[2] = d2;  o[3] = d3;
-    o[4] = d4;  o[5] = d5;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;  ohi[2] = d2;  ohi[3] = d3;
+    ohi[4] = d4;  ohi[5] = d5;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 6);
 }
-static inline WDataOutP VL_CONSTHI_W_7X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_7X(int obits, int lsb, WDataOutP o,
                                        EData d6, EData d5, EData d4,
                                        EData d3, EData d2, EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;  o[2] = d2;  o[3] = d3;
-    o[4] = d4;  o[5] = d5;  o[6] = d6;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;  ohi[2] = d2;  ohi[3] = d3;
+    ohi[4] = d4;  ohi[5] = d5;  ohi[6] = d6;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 7);
 }
-static inline WDataOutP VL_CONSTHI_W_8X(int obits, int lsb, WDataOutP obase,
+static inline WDataOutP VL_CONSTHI_W_8X(int obits, int lsb, WDataOutP o,
                                        EData d7, EData d6, EData d5, EData d4,
                                        EData d3, EData d2, EData d1, EData d0) VL_MT_SAFE {
-    WDataOutP o = obase + VL_WORDS_I(lsb);
-    o[0] = d0;  o[1] = d1;  o[2] = d2;  o[3] = d3;
-    o[4] = d4;  o[5] = d5;  o[6] = d6;  o[7] = d7;
+    WDataOutP ohi = o + VL_WORDS_I(lsb);
+    ohi[0] = d0;  ohi[1] = d1;  ohi[2] = d2;  ohi[3] = d3;
+    ohi[4] = d4;  ohi[5] = d5;  ohi[6] = d6;  ohi[7] = d7;
    VL_C_END_(obits, VL_WORDS_I(lsb) + 8);
 }

@@ -2164,6 +2223,9 @@ inline IData VL_CMP_NN(const std::string& lhs, const std::string& rhs, bool igno
 }

 extern IData VL_ATOI_N(const std::string& str, int base) VL_PURE;
+extern IData VL_NTOI_I(int obits, const std::string& str) VL_PURE;
+extern QData VL_NTOI_Q(int obits, const std::string& str) VL_PURE;
+extern void VL_NTOI_W(int obits, WDataOutP owp, const std::string& str) VL_PURE;

 extern IData VL_FGETS_NI(std::string& dest, IData fpi) VL_MT_SAFE;

@@ -2182,6 +2244,7 @@ extern IData VL_DIST_UNIFORM(IData& seedr, IData ustart, IData uend) VL_MT_SAFE;
 // Conversion functions

 extern std::string VL_CVT_PACK_STR_NW(int lwords, const WDataInP lwp) VL_PURE;
+extern std::string VL_CVT_PACK_STR_ND(const VlQueue<std::string>& q) VL_PURE;
 inline std::string VL_CVT_PACK_STR_NQ(QData lhs) VL_PURE {
    VlWide<VL_WQ_WORDS_E> lw;
    VL_SET_WQ(lw, lhs);
@@ -2198,10 +2261,10 @@ inline std::string VL_CONCATN_NNN(const std::string& lhs, const std::string& rhs
    return lhs + rhs;
 }
 inline std::string VL_REPLICATEN_NNQ(const std::string& lhs, IData rep) VL_PURE {
-    std::string out;
-    out.reserve(lhs.length() * rep);
-    for (unsigned times = 0; times < rep; ++times) out += lhs;
-    return out;
+    std::string result;
+    result.reserve(lhs.length() * rep);
+    for (unsigned times = 0; times < rep; ++times) result += lhs;
+    return result;
 }
 inline std::string VL_REPLICATEN_NNI(const std::string& lhs, IData rep) VL_PURE {
    return VL_REPLICATEN_NNQ(lhs, rep);
@@ -2212,6 +2275,7 @@ extern std::string VL_TOLOWER_NN(const std::string& ld) VL_PURE;
 extern std::string VL_TOUPPER_NN(const std::string& ld) VL_PURE;

 extern IData VL_FERROR_IN(IData fpi, std::string& outputr) VL_MT_SAFE;
+extern IData VL_FERROR_IW(IData fpi, int obits, WDataOutP outwp) VL_MT_SAFE;
 extern IData VL_FOPEN_NN(const std::string& filename, const std::string& mode) VL_MT_SAFE;
 extern IData VL_FOPEN_MCD_N(const std::string& filename) VL_MT_SAFE;
 extern void VL_READMEM_N(bool hex, int bits, QData depth, int array_lsb,
--- a/Show More
+++ b/Show More