verilator/docs
Thomas Santerre bd6b9161dc
Optimize bit-scan loops into $mostsetbitp1 / $countones (#7822)
Recognize the common single-bit scan loop idioms in V3Unroll (before it
unrolls) and lower them to bit-reduction primitives, replacing a literal
W-iteration loop with one intrinsic-backed expression:

  target=0; for (i=0;i<W;i++) if (vec[i]) target = i + 1;      -> $mostsetbitp1(vec)
  target=0; for (i=0;i<W;i++) if (vec[i]) target = target + 1; -> $countones(vec)

The leading-one form lowers to a new AstMostSetBitP1 node, emitted as
VL_MOSTSETBITP1_{I,Q,W}; those runtime helpers now use __builtin_clz where
available (same pattern as VL_REDXOR's __builtin_parity), with the existing
bit scan as fallback.  The count-ones form reuses AstCountOnes ($countones,
popcount); as the DFG requires a 32-bit countones result it is built at 32
bits and narrowed to the accumulator width with a select.

Matching is structural to stay sound: the index must start at 0, increment
by exactly 1, and scan all W==width(vec) bits via a single 1-bit select of a
distinct vector, with the target pre-zeroed and no else branch.  The loop
bound is accepted as a strict ascending 'idx < W' written either way and
signed or unsigned (Gt/GtS/Lt/LtS).  Gated by -fbit-scan-loops (on at -O).

Adds t_bit_scan_loops (I/Q/W, count-ones and unsigned-index positives;
step-2, start-1, idx*2+1, vec[idx+1], target=idx and W!=width negatives, all
self-checked and asserted via --stats not to lower) plus t_bit_scan_loops_off
for the disable flag.

Motivated by a transformer inference design whose 80-bit leading-one detector
ran every cycle (~37% of runtime); the lowering is worth ~39% there.
2026-06-24 10:43:05 +01:00
..
_static Fix Codacy warnings. No functional change. 2021-07-07 19:42:49 -04:00
bin Add SPDX copyright identifiers, and get 'reuse' clean. No functional change. 2026-01-26 20:24:34 -05:00
gen Support NBAs in initial blocks (#7754) 2026-06-20 17:23:05 -04:00
guide Optimize bit-scan loops into $mostsetbitp1 / $countones (#7822) 2026-06-24 10:43:05 +01:00
.gitignore Spelling fixes. 2022-05-14 16:12:57 -04:00
AGENTS.md CI: Autoformat markdown files 2026-06-15 17:44:50 -04:00
CONTRIBUTING.rst Commentary: Use standard multiline rst comments, other cleanups 2026-06-18 21:58:01 -04:00
CONTRIBUTORS Optimize bit-scan loops into $mostsetbitp1 / $countones (#7822) 2026-06-24 10:43:05 +01:00
Makefile Add SPDX copyright identifiers, and get 'reuse' clean. No functional change. 2026-01-26 20:24:34 -05:00
README.rst Commentary: Use standard multiline rst comments, other cleanups 2026-06-18 21:58:01 -04:00
internals.rst Commentary: Make RST documents round-trip clean. No output change intended. 2026-06-21 10:15:47 -04:00
security.rst Commentary: Use standard multiline rst comments, other cleanups 2026-06-18 21:58:01 -04:00
spelling.txt Commentary: Changes update 2026-06-18 21:56:02 -04:00
verilated.dox Cleanup missing copyrights and those on simply copied files. No functional change. 2023-01-20 20:42:30 -05:00

README.rst

..
   SPDX-FileCopyrightText: 2003-2026 Wilson Snyder
   SPDX-License-Identifier: LGPL-3.0-only OR Artistic-2.0

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>`_