Commit Graph

13 Commits

Author SHA1 Message Date
Geza Lore d33a81d32a Optimize commutative vertex operands in Dfg for better combining 2026-03-25 08:09:37 +00:00
Geza Lore 5834f22944
Optimize logic and variable removal early in Dfg (#7081)
After conversion of Ast to Dfg, but before synthesizing AstAlways into
primitives, run a pass to remove variables that are not observable, and
all logic that only computes such variables. This can get rid of a lot
of content early so we don't build redundant Dfgs, and also enables
synthesizing always blocks that use temporaries only in some branches,
which will come in a follow up.
2026-02-17 08:28:06 +00:00
Geza Lore be27811a10
Optimize: Remove redundant variables during Dfg Peephole pass (#7076)
Many rules in the Dfg Peephole pass check if a node has more than one
sinks. Redundant variables that will ultimately be removed can prevent
these from matching. Remove such variables during the Peeophole pass
itself to enable more matches.
2026-02-14 12:33:20 +00:00
Wilson Snyder 7c6c6a684b Add SPDX copyright identifiers, and get 'reuse' clean. No functional change. 2026-01-26 20:24:34 -05:00
Geza Lore af1be26b96
Fix O(n*2) analysis in const-bit-op-tree (#6791)
Note this might miss some cases where a sub-tree within an And/Or/Xor
tree is optimizeable, but not the whole tree, but in practice this seems
to work better than the alternative of keeping a set of failed nodes and
bail early.
2025-12-11 14:32:25 +00:00
Geza Lore e494cf22a4
Add DfgPeephole patterns (#6149) 2025-07-07 16:25:29 +01:00
Geza Lore 7a3f1f16ca
Optimize DFG before V3Gate (#6141) 2025-07-01 17:55:08 -04:00
Yutetsu TAKATSUKASA 100e3d7702
Fix const-bit-op-tree with single-bit masks (#5993) (#5998) 2025-05-10 06:01:15 -04:00
Wilson Snyder abc46db3c0 Tests: Cleanup some .cpp paths. No test functional change indended. 2025-04-25 20:57:11 -04:00
Geza Lore d9701e6406
Automatically split some packed variables (#5843)
This patch adds a heuristic to V3SplitVar, and it attempts to split up
packed variables that are only referenced via constant index,
non-overlapping bit/range selects. This can eliminate some UNOPTFLAT cases.
2025-03-09 10:31:01 -04:00
Geza Lore 03bd1bfc63
Move Concat balancing from DFG to FuncOpt (#5602)
This means it applies more widely, e.g. inside sequential logic.
2024-11-10 17:23:11 +00:00
Geza Lore f073b278f9
Balance concatenations in DFG (#5598)
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).
2024-11-09 18:14:19 +00:00
Wilson Snyder 2c52119d6c Tests: Rename some tests into proper groups. No functional change. 2024-09-21 10:04:58 -04:00