https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115118
Filip Kastl <pheeck at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[15 Regression] 5-13% |[15 Regression] 5-13%
|slowdown of 470.lbm on zen4 |slowdown of 470.lbm on zen4
| |since
| |r15-327-gf4b86ab09dfe25
Keywords|needs-bisection |
CC| |aagarwa at gcc dot gnu.org
--- Comment #4 from Filip Kastl <pheeck at gcc dot gnu.org> ---
Yep, r15-327-gf4b86ab09dfe25 increases exec time by 4,5% (on the machine where
I previously measured 5%). Thanks Andrew. Cc-ing the author.
commit f4b86ab09dfe258c4780fcc7567ca8a275c96e7a
Author: Ajit Kumar Agarwal <[email protected]>
AuthorDate: Wed May 8 09:21:27 2024 -0500
Commit: Ajit Kumar Agarwal <[email protected]>
CommitDate: Wed May 8 09:23:45 2024 -0500
tree-ssa-sink: Improve code sinking pass
Currently, code sinking will sink code at the use points with loop having
same
nesting depth. The following patch improves code sinking by placing the
sunk
code in begining of the block after the labels.
2024-05-08 Ajit Kumar Agarwal <[email protected]>
gcc/ChangeLog:
PR tree-optimization/81953
* tree-ssa-sink.cc (statement_sink_location):Sink statements at
the begining of the basic block after labels.
gcc/testsuite/ChangeLog:
PR tree-optimization/81953
* gcc.dg/tree-ssa/ssa-sink-21.c: New test.
