public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r11-7099] tree-optimization/98855 - fix some vectorizer cost issues
@ 2021-02-04 12:02 Richard Biener
0 siblings, 0 replies; only message in thread
From: Richard Biener @ 2021-02-04 12:02 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:7002a33d1ba81e4577d965fb9daaee146b31faa8
commit r11-7099-g7002a33d1ba81e4577d965fb9daaee146b31faa8
Author: Richard Biener <rguenther@suse.de>
Date: Thu Feb 4 12:08:47 2021 +0100
tree-optimization/98855 - fix some vectorizer cost issues
This fixes us not costing vectorized bswap for SLP as well as
avoiding biasing to the vectorized side when costing single-argument
PHIs. Instead we assume coalescing here and cost them with zero cost
for both the scalar and vectorized code.
This doesn't fix the PR on its own.
2021-02-04 Richard Biener <rguenther@suse.de>
PR tree-optimization/98855
* tree-vect-loop.c (vectorizable_phi): Do not cost
single-argument PHIs.
* tree-vect-slp.c (vect_bb_slp_scalar_cost): Likewise.
* tree-vect-stmts.c (vectorizable_bswap): Also perform
costing for SLP operation.
Diff:
---
gcc/tree-vect-loop.c | 8 ++++++--
gcc/tree-vect-slp.c | 7 +++++++
gcc/tree-vect-stmts.c | 13 ++++++-------
3 files changed, 19 insertions(+), 9 deletions(-)
diff --git a/gcc/tree-vect-loop.c b/gcc/tree-vect-loop.c
index 200ed27b32e..f1f16e3b1b9 100644
--- a/gcc/tree-vect-loop.c
+++ b/gcc/tree-vect-loop.c
@@ -7777,8 +7777,12 @@ vectorizable_phi (vec_info *,
"incompatible vector types for invariants\n");
return false;
}
- record_stmt_cost (cost_vec, SLP_TREE_NUMBER_OF_VEC_STMTS (slp_node),
- vector_stmt, stmt_info, vectype, 0, vect_body);
+ /* For single-argument PHIs assume coalescing which means zero cost
+ for the scalar and the vector PHIs. This avoids artificially
+ favoring the vector path (but may pessimize it in some cases). */
+ if (gimple_phi_num_args (as_a <gphi *> (stmt_info->stmt)) > 1)
+ record_stmt_cost (cost_vec, SLP_TREE_NUMBER_OF_VEC_STMTS (slp_node),
+ vector_stmt, stmt_info, vectype, 0, vect_body);
STMT_VINFO_TYPE (stmt_info) = phi_info_type;
return true;
}
diff --git a/gcc/tree-vect-slp.c b/gcc/tree-vect-slp.c
index 10b876ff5ed..2305bbdec3a 100644
--- a/gcc/tree-vect-slp.c
+++ b/gcc/tree-vect-slp.c
@@ -4297,6 +4297,13 @@ vect_bb_slp_scalar_cost (vec_info *vinfo,
}
else if (vect_nop_conversion_p (orig_stmt_info))
continue;
+ /* For single-argument PHIs assume coalescing which means zero cost
+ for the scalar and the vector PHIs. This avoids artificially
+ favoring the vector path (but may pessimize it in some cases). */
+ else if (is_a <gphi *> (orig_stmt_info->stmt)
+ && gimple_phi_num_args
+ (as_a <gphi *> (orig_stmt_info->stmt)) == 1)
+ continue;
else
kind = scalar_stmt;
record_stmt_cost (cost_vec, 1, kind, orig_stmt_info,
diff --git a/gcc/tree-vect-stmts.c b/gcc/tree-vect-stmts.c
index 5eb7b2d1d6e..0bc1cb1c5b4 100644
--- a/gcc/tree-vect-stmts.c
+++ b/gcc/tree-vect-stmts.c
@@ -3002,13 +3002,12 @@ vectorizable_bswap (vec_info *vinfo,
STMT_VINFO_TYPE (stmt_info) = call_vec_info_type;
DUMP_VECT_SCOPE ("vectorizable_bswap");
- if (! slp_node)
- {
- record_stmt_cost (cost_vec,
- 1, vector_stmt, stmt_info, 0, vect_prologue);
- record_stmt_cost (cost_vec,
- ncopies, vec_perm, stmt_info, 0, vect_body);
- }
+ record_stmt_cost (cost_vec,
+ 1, vector_stmt, stmt_info, 0, vect_prologue);
+ record_stmt_cost (cost_vec,
+ slp_node
+ ? SLP_TREE_NUMBER_OF_VEC_STMTS (slp_node) : ncopies,
+ vec_perm, stmt_info, 0, vect_body);
return true;
}
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2021-02-04 12:02 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-04 12:02 [gcc r11-7099] tree-optimization/98855 - fix some vectorizer cost issues Richard Biener
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).