Mirrors
/
gcc
mirror of https://github.com/gcc-mirror/gcc.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

sra: Do not leave work for DSE (that it can sometimes not perform) 

When looking again at the g++.dg/tree-ssa/pr109849.C testcase we
discovered that it generates terrible store-to-load forwarding stalls
because SRA was leaving behind aggregate loads but all the stores were
by scalar parts and DSE failed to remove the useless load.  SRA has
all the knowledge to remove the statement even now, so this small
patch makes it do so.

With this patch, the g++.dg/tree-ssa/pr109849.C micro-benchmark runs 9
times faster (on an AMD EPYC 75F3 machine).

gcc/ChangeLog:

2024-04-18  Martin Jambor  <mjambor@suse.cz>

	* tree-sra.cc (sra_modify_assign): Remove the original statement
	also when dealing with a store to a fully covered aggregate from a
	non-candidate.

gcc/testsuite/ChangeLog:

2024-04-23  Martin Jambor  <mjambor@suse.cz>

	* g++.dg/tree-ssa/pr109849.C: Also check that the aggeegate store
	to cur disappears.
	* gcc.dg/tree-ssa/ssa-dse-26.c: Instead of relying on DSE,
	check that the unwanted stores were removed at early SRA time.
pull/52/merge
Martin Jambor 2024-05-09 16:39:44 +02:00
parent 2790195500
commit f6743695b4
No known key found for this signature in database
GPG Key ID: BF63C1BC3FA43540
3 changed files with 17 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
gcc/testsuite/g++.dg/tree-ssa/pr109849.C
View File

@ -1,5 +1,5 @@
/* { dg-do compile } */
/* { dg-options "-O2 -fdump-tree-sra" } */
/* { dg-options "-O2 -fdump-tree-sra -fdump-tree-optimized" } */
#include <vector>
typedef unsigned int uint32_t;
@ -29,3 +29,4 @@ main()
}
/* { dg-final { scan-tree-dump "Created a replacement for stack offset" "sra"} } */
/* { dg-final { scan-tree-dump-not "cur = MEM" "optimized"} } */

6
gcc/testsuite/gcc.dg/tree-ssa/ssa-dse-26.c
View File

@ -1,5 +1,5 @@
/* { dg-do compile } */
/* { dg-options "-O2 -fdump-tree-dse1-details -fno-short-enums -fno-tree-fre" } */
/* { dg-options "-O2 -fdump-tree-esra -fno-short-enums -fno-tree-fre" } */
/* { dg-skip-if "we want a BIT_FIELD_REF from fold_truth_andor" { ! lp64 } } */
/* { dg-skip-if "temporary variable names are not x and y" { mmix-knuth-mmixware } } */
@ -31,5 +31,5 @@ constraint_equal (struct constraint a, struct constraint b)
&& constraint_expr_equal (a.rhs, b.rhs);
}
/* { dg-final { scan-tree-dump-times "Deleted dead store: x = " 2 "dse1" } } */
/* { dg-final { scan-tree-dump-times "Deleted dead store: y = " 2 "dse1" } } */
/* { dg-final { scan-tree-dump-not "x = " "esra" } } */
/* { dg-final { scan-tree-dump-not "y = " "esra" } } */

14
gcc/tree-sra.cc
View File

@ -4854,8 +4854,18 @@ sra_modify_assign (gimple *stmt, gimple_stmt_iterator *gsi)
But use the RHS aggregate to load from to expose more
optimization opportunities. */
if (access_has_children_p (lacc))
generate_subtree_copies (lacc->first_child, rhs, lacc->offset,
0, 0, gsi, true, true, loc);
{
generate_subtree_copies (lacc->first_child, rhs, lacc->offset,
0, 0, gsi, true, true, loc);
if (lacc->grp_covered)
{
unlink_stmt_vdef (stmt);
gsi_remove (& orig_gsi, true);
release_defs (stmt);
sra_stats.deleted++;
return SRA_AM_REMOVED;
}
}
}
return SRA_AM_NONE;