The AstIf nodes conditional on events being triggered used to be created
in V3Clock. Now it is in V3Sched*, in order to avoid having to pass
AstActive in CFunc or MTask bodies. No functional change intended, some
improved optimization due to simplifying timing triggers that were
previously missed, also fixes what seems like a bug in the original
timing commit code.
sched_forks.tree used to be dumped before sched.tree, while it's
basically after, so move transformForks in to a separate pass. Also
extract inlined visitors in V3SchedTiming.
I stared this because the emitted makefiles for hierarchical verilation
were non-deterministic (iterating unordered_map indexed by pointers).
Then I realized that the V3HierPlan is just a dependency graph encoded
in a slightly idiosyncratic way. We do have a data structure to use for
that instead.
With that the output should always be deterministic + have nicer dumps.
This patch implements #6480. All loop statements are represented using
AstLoop and AstLoopTest.
This necessitates rework of the loop unroller to handle loops of
arbitrary form. To enable this, I have split the old unroller used for
'generate for' statements and moved it into V3Param, and subsequently
rewrote V3Unroll to handle the new representation. V3Unroll can now
unroll more complex loops, including with loop conditions containing
multiple variable references or inlined functions.
Handling the more generic code also requires some restrictions. If a
loop contains any of the following, it cannot be unrolled:
- A timing control that might suspend the loop
- A non-inlined call to a non-pure function
These constructs can change the values of variables in the loop, so are
generally not safe to unroll if they are present. (We could still unroll
if all the variables needed for unrolling are automatic, however we
don't do that right now.)
These restrictions seem ok in the benchmark suite, where the new
unroller can generally unroll many more loops than before.
Internals: Refactor generate construct Ast handling (#6280)
We introduce AstNodeGen, the common base class of AstGenBlock,
AstGenCase, AstGenFor, and AstGenIf, which together represent all SV
generate constructs. Subsequently remove AstNodeFor, AstNodeCase
(AstCase is now directly derived from AstNodeStmt) and adjust internals
to work on the new representation.
Output is identical modulo hashes do to changed AstNode type ids, no
functional change intended.
Step towards #6280.
Rename AstAssignAlias to AstAlias and make it derive from AstNode
instead of AstNodeStmt.
Replace AstAlias with AstAssignW in V3LinkDot::linkDotScope, which is
the last place we need to be aware of the alias construct. Using
AstAssignW dowstream enables further optimization while preserving the
same functionality.
Replace std::exit with v3Global.exit, and make V3Error::vlAbort call
v3Global.shutdown. This gives us an opportunity to release resources to
facilitate leak checking even when exiting early on an error.
Note we still don't release most resources by default without
VL_LEAK_CHECKS, so there is no behaviour change there.
Add a new pass to split up (recursively):
foo = {l, r};
into the following, with the right indices, iff the concatenation
straddles a wide word boundary.
foo[_:_] = r;
foo[_:_] = l;
This eliminates more wide temporaries.
Another 23% speedup on VeeR EH2 high_perf. Also brings the predicted
stack size from 8M to 40k.
The DFG peephole pass converts all associative trees into right leaning,
which is good for simplifying pattern recognition, but can lead to an
excessive amount of wide intermediate results being constructed for
right leaning concatenations.
Add a new pass to balance concatenation trees by trying to:
- Create VL_EDATASIZE (32-bit) sub-terms, so words can then be packed
easily afterwards
- Try to ensure the operands of a concat are roughly the same width
within a concatenation tree. This does not yield the shortest tree,
but it ensures it has many sub-nodes that are small enough to fit into
machine registers.
This can eliminate a lot of wide intermediate results, which would need
temporaries, and also increases ILP within sub-expressions (assuming the
C compiler can't figure that out itself).
This is over 2x run-time speedup on the high_perf configuration of
VeeR EH2 (which you could arguably also get with -fno-dfg, but oh well).
Instead of carrying around MTask affinity from scheduling, compute it in
V3VariableOrder (where it is used), by tracing through the code. This
simplifies some code and has the benefit of handling variables
introduced after scheduling. It's worth a few % speed at run-time, and
the new implementation of V3VariableOrder is slightly more efficient,
though the speed/space is still dominated by the TSP sort.
V3Partition used to contain 2 conceptually separate set of algorithms
- The MTask partitioning/coarsening algorithm used by V3Order. This has
been moved to V3OrderParallel.cpp
- The lowering of AstExecGraph into per thread functions by packing
tasks into threads and creating additional code
(V3Partition::finalize). This has been moved to the new
V3ExecGraph.cpp
This patch is just code movement/rename with minimal fixes required to
do so.
With --stats, we will print DFG pattern combinations, one per line, as
S-expressions to new stat files, together with their frequency, to aid
discovery of new peephole patterns.
Wilson Snyder
Bartłomiej Chmiel
Geza Lore
Krzysztof Sychla
Andrew Voznytsa
Arkadiusz Kozdra
Szymon Gizler
Szymon Gizler
Krzysztof Bieganski