- Improve flag pruning heuristic
- Set all trace activity flags in slow code. This in turns enables us
to remove checking the slow flag on the fast path.
Firstly, we always use a byte array for fine grained activity flags
instead of a bit vector (we used to use a byte array only if we had
parallel mtasks). The byte vector can be set more cheaply in eval,
closing about 1/3 of the gap in performance between compiling with
or without --trace on SweRV EH1. The speed of tracing itself is not
measurably different.
Secondly, we prune the activity tracking such that if a set of activity
flag combinations only guard a small number of signals, we will turn
those signals into awayls traced signals. This avoids code which
sometimes tests dozens of activity flags just to subsequently check one
signal and dump it if it's value changed. We can just check the signal
state straight instead, and not bother with the flags. This removes
about 30% of activity flags in SweRV EH1, and makes both single threaded
VCD and FST tracing 8-9% faster.
The main goal of this patch is to enable splitting the full and
incremental tracing functions into multiple functions, which can then be
run in parallel at a later stage. It also simplifies further
experimentation as all of the interesting trace code construction now
happens in V3Trace. No functional change is intended by this patch, but
there are some implementation changes in the generated code.
Highlights:
- Pass symbol table directly to trace callbacks for simplicity.
- A new traceRegister function is generated which adds each trace
function as an individual callback, which means we can have multiple
callbacks for each trace function type.
- A new traceCleanup function is generated which clears the activity
flags, as the trace callbacks might be implemented as multiple functions.
- Re-worked sub-function handling so there is no separate sub-function
for each trace activity class. Sub-functions are generate when required
by splitting.
- traceFull/traceChg are now created in V3Trace rather than V3TraceDecl,
this requires carrying the trace value tree in TraceDecl until it
reaches V3Trace where the TraceInc nodes are created (previously a
TraceInc was also created in V3TraceDecl which carries the value).
- Allow arbitrary number of open array dimensions, not just 3. Note
right now this only works with the array querying functions specified
in IEEE 1800-2017 H.12.2
- Issue error when passing dynamic array or queue as DPI open array
(currently unsupported)
- Also tweaked AstVar::vlArgTypeRecurse, which should now error or fail
for unsupported types.
We used to iterate the m_callMmap std::multimap by getting limit
iterators from equal_range, but we also modify the same map in the loop
which invalidates those limit iterators. Note this only caused actual
problems if the new AstCCall inserted via 'addCall' in the loop had a
memory address (which is used as the key) which fell into the range
returned by equal_range, so was pretty hard to trigger.
- Issue an error when --build is used together with --make
- When given --build, always use GNU Make to perform the build
- Update documentation (examples were good as they were)
- Remove the broken t_flag_build_cmake test
Fixes#2280
Use SIMD intrinsics to render VCD traces.
I have measured 10-40% single threaded performance increase with VCD
tracing on SweRV EH1 and lowRISC Ibex using SSE2 intrinsics to render
the trace. Also helps a tiny bit with FST, but now almost all of the FST
overhead is in the FST library.
I have reworked the tracing routines to use more precisely sized
arguments. The nice thing about this is that the performance without the
intrinsics is pretty much the same as it was before, as we do at most 2x
as much work as necessary, but in exchange there are no data dependent
branches at all.
Convert trace buffer to 32-bit entries, rather than a union containing a
pointer type. Also tweaked trace entry layouts for a bit more
performance. This gains another 10% on SweRV EH1 CoreMark.
- Change templated trace routines to branch table.
Removed templating from trace chgBus and fullBus and replaced them with
a branch table like the other there is a very small (< 1%) penalty for
this on SwerRV EH1 CoreMark, but this is less than the variability of
disk IO so it's worth it to keep the code simpler and smaller.
- Prefetch VCD suffix buffer at the top of emit*
- Increase ILP in VCD emit* routines
- Use a 64-bit unaligned store to emit the VCD suffix (on x86 only)
The performance difference with these is very small, but the changes
hopefully make this code more performance-portable across various
micro-architectures.
We used to include a .cpp file on the link line for the shared library,
which was ignored, but generated a .d file for the .so which contained
the header files required by the .cpp file. This then caused a rebuild
where we included the .d in verilated.mk to included in the .h headers
among the prerequisites of the .so, yielding a clang error about treating
.h files as c++-header rather than c-header... Long story short, we don't
do that anymore. This used t cause t_a4_examples to fail on occasion.
Note there is no need for a separate compilation rule for the
<--protect-lib>.cpp, as it will jsut pick up the standard OPT_FAST rule.
The --trace-threads option can now be used to perform tracing on a
thread separate from the main thread when using VCD tracing (with
--trace-threads 1). For FST tracing --trace-threads can be 1 or 2, and
--trace-fst --trace-threads 1 is the same a what --trace-fst-threads
used to be (which is now deprecated).
Performance numbers on SweRV EH1 CoreMark, clang 6.0.0, Intel i7-3770 @
3.40GHz, IO to ramdisk, with numactl set to schedule threads on different
physical cores. Relative speedup:
--trace -> --trace --trace-threads 1 +22%
--trace-fst -> --trace-fst --trace-threads 1 +38% (as --trace-fst-thread)
--trace-fst -> --trace-fst --trace-threads 2 +93%
Speed relative to --trace with no threaded tracing:
--trace 1.00 x
--trace --trace-threads 1 0.82 x
--trace-fst 1.79 x
--trace-fst --trace-threads 1 1.23 x
--trace-fst --trace-threads 2 0.87 x
This means FST tracing with 2 extra threads is now faster than single
threaded VCD tracing, and is on par with threaded VCD tracing. You do
pay for it in total compute though as --trace-fst --trace-threads 2 uses
about 240% CPU vs 150% for --trace-fst --trace-threads 1, and 155% for
--trace --trace threads 1. Still for interactive use it should be
helpful with large designs.
Includes `timescale, $printtimescale, $timeformat.
VL_TIME_MULTIPLIER, VL_TIME_PRECISION, VL_TIME_UNIT have been removed
and the time precision must now match the SystemC time precision.
To get closer behavior to older versions, use e.g. --timescale-override
"1ps/1ps".
* Improve tracing performance.
Various tactics used to improve performance of both VCD and FST tracing:
- Both: Change tracing functions to templates to take variable widths as
template parameters. For VCD, subsequently specialize these to the
values used by Verilator. This avoids redundant instructions and hard
to predict branches.
- Both: Check for value changes via direct pointer access into the
previous signal value buffer. This eliminates a lot of simple pointer
arithmetic instructions form the tracing code.
- Both: Verilator provides clean input, no need to mask out used bits.
- VCD: pre-compute identifier codes and use memory copy instead of
re-computing them every time a code is emitted. This saves a lot of
instructions and hard to predict branches. The added D-cache misses
are cheaper than the removed branches/instructions.
- VCD: re-write the routines emitting the changes to be more efficient.
- FST: Use previous signal value buffer the same way as the VCD tracing
code, and only call the FST API when a change is detected.
Performance as measured on SweRV EH1, with the pre-canned CoreMark
benchmark running from DCCM/ICCM, clang 6.0.0, Intel i7-3770 @ 3.40GHz,
and IO to ramdisk:
+--------------+---------------+----------------------+
| VCD | FST | FST separate thread |
| (--trace) | (--trace-fst) | (--trace-fst-thread) |
------------+-----------------------------------------------------+
Before | 30.2 s | 121.1 s | 69.8 s |
============+==============+===============+======================+
After | 24.7 s | 45.7 s | 32.4 s |
------------+--------------+---------------+----------------------+
Speedup | 22 % | 256 % | 215 % |
------------+--------------+---------------+----------------------+
Rel. to VCD | 1 x | 1.85 x | 1.31 x |
------------+--------------+---------------+----------------------+
In addition, FST trace size for the above reduced by 48%.
This looks like a bits/bytes bug. The affected m_codeInc member
determines how many 32-bit words to allocate in a buffer used to store
previous values of the signal, but this was off by a factor of 8, so
we used to use too much memory.
SweRV VCD tracing speed +6.5% (excluding IO, clang 6.0), due mainly to
reduced D cache misses.
Wilson Snyder
Geza Lore
Stephen Henry
Huang Rui
Yossi Nivin
Yutetsu TAKATSUKASA
James Hanlon