public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r12-9902] Disable generation of scalar modulo instructions.
@ 2023-10-02 20:06 Pat Haugen
0 siblings, 0 replies; only message in thread
From: Pat Haugen @ 2023-10-02 20:06 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:65d89dd9ae151a19fd403e1fbcb1b37b942d2932
commit r12-9902-g65d89dd9ae151a19fd403e1fbcb1b37b942d2932
Author: Pat Haugen <pthaugen@linux.ibm.com>
Date: Tue Sep 19 13:19:59 2023 -0500
Disable generation of scalar modulo instructions.
It was recently discovered that the scalar modulo instructions can suffer
noticeable performance issues for certain input values. This patch disables
their generation since the equivalent div/mul/sub sequence does not suffer
the same problem.
gcc/
* config/rs6000/rs6000.cc (rs6000_rtx_costs): Check whether the
modulo instruction is disabled.
* config/rs6000/rs6000.h (RS6000_DISABLE_SCALAR_MODULO): New.
* config/rs6000/rs6000.md (mod<mode>3, *mod<mode>3): Check it.
(define_expand umod<mode>3): New.
(define_insn umod<mode>3): Rename to *umod<mode>3 and check if the modulo
instruction is disabled.
(umodti3, modti3): Check if the modulo instruction is disabled.
gcc/testsuite/
* gcc.target/powerpc/clone1.c: Add xfails.
* gcc.target/powerpc/clone3.c: Likewise.
* gcc.target/powerpc/mod-1.c: Update scan strings and add xfails.
* gcc.target/powerpc/mod-2.c: Likewise.
* gcc.target/powerpc/p10-vdivq-vmodq.c: Add xfails.
(cherry picked from commit 58ab38213b979811d314f68e3f455c28a1d44140)
Diff:
---
gcc/config/rs6000/rs6000.cc | 4 ++-
gcc/config/rs6000/rs6000.h | 6 ++++
gcc/config/rs6000/rs6000.md | 40 +++++++++++++++++++---
gcc/testsuite/gcc.target/powerpc/clone1.c | 7 ++--
gcc/testsuite/gcc.target/powerpc/clone3.c | 7 ++--
gcc/testsuite/gcc.target/powerpc/mod-1.c | 21 ++++++------
gcc/testsuite/gcc.target/powerpc/mod-2.c | 11 +++---
gcc/testsuite/gcc.target/powerpc/p10-vdivq-vmodq.c | 5 +--
8 files changed, 72 insertions(+), 29 deletions(-)
diff --git a/gcc/config/rs6000/rs6000.cc b/gcc/config/rs6000/rs6000.cc
index dd263fbf9fe..9fedf248f24 100644
--- a/gcc/config/rs6000/rs6000.cc
+++ b/gcc/config/rs6000/rs6000.cc
@@ -21907,7 +21907,9 @@ rs6000_rtx_costs (rtx x, machine_mode mode, int outer_code,
*total = rs6000_cost->divsi;
}
/* Add in shift and subtract for MOD unless we have a mod instruction. */
- if (!TARGET_MODULO && (code == MOD || code == UMOD))
+ if ((!TARGET_MODULO
+ || (RS6000_DISABLE_SCALAR_MODULO && SCALAR_INT_MODE_P (mode)))
+ && (code == MOD || code == UMOD))
*total += COSTS_N_INSNS (2);
return false;
diff --git a/gcc/config/rs6000/rs6000.h b/gcc/config/rs6000/rs6000.h
index 5c886d909aa..9577e715d67 100644
--- a/gcc/config/rs6000/rs6000.h
+++ b/gcc/config/rs6000/rs6000.h
@@ -2632,3 +2632,9 @@ while (0)
rs6000_asm_output_opcode (STREAM); \
} \
while (0)
+
+/* Disable generation of scalar modulo instructions due to performance issues
+ with certain input values. This can be removed in the future when the
+ issues have been resolved. */
+#define RS6000_DISABLE_SCALAR_MODULO 1
+
diff --git a/gcc/config/rs6000/rs6000.md b/gcc/config/rs6000/rs6000.md
index 9d5e0244d81..7aee059a39f 100644
--- a/gcc/config/rs6000/rs6000.md
+++ b/gcc/config/rs6000/rs6000.md
@@ -3359,6 +3359,17 @@
FAIL;
operands[2] = force_reg (<MODE>mode, operands[2]);
+
+ if (RS6000_DISABLE_SCALAR_MODULO)
+ {
+ temp1 = gen_reg_rtx (<MODE>mode);
+ temp2 = gen_reg_rtx (<MODE>mode);
+
+ emit_insn (gen_div<mode>3 (temp1, operands[1], operands[2]));
+ emit_insn (gen_mul<mode>3 (temp2, temp1, operands[2]));
+ emit_insn (gen_sub<mode>3 (operands[0], operands[1], temp2));
+ DONE;
+ }
}
else
{
@@ -3378,17 +3389,36 @@
[(set (match_operand:GPR 0 "gpc_reg_operand" "=&r")
(mod:GPR (match_operand:GPR 1 "gpc_reg_operand" "r")
(match_operand:GPR 2 "gpc_reg_operand" "r")))]
- "TARGET_MODULO"
+ "TARGET_MODULO && !RS6000_DISABLE_SCALAR_MODULO"
"mods<wd> %0,%1,%2"
[(set_attr "type" "div")
(set_attr "size" "<bits>")])
+;; This define_expand can be removed when RS6000_DISABLE_SCALAR_MODULO is
+;; removed.
+(define_expand "umod<mode>3"
+ [(set (match_operand:GPR 0 "gpc_reg_operand")
+ (umod:GPR (match_operand:GPR 1 "gpc_reg_operand")
+ (match_operand:GPR 2 "gpc_reg_operand")))]
+ "TARGET_MODULO"
+{
+ if (RS6000_DISABLE_SCALAR_MODULO)
+ {
+ rtx temp1 = gen_reg_rtx (<MODE>mode);
+ rtx temp2 = gen_reg_rtx (<MODE>mode);
+
+ emit_insn (gen_udiv<mode>3 (temp1, operands[1], operands[2]));
+ emit_insn (gen_mul<mode>3 (temp2, temp1, operands[2]));
+ emit_insn (gen_sub<mode>3 (operands[0], operands[1], temp2));
+ DONE;
+ }
+})
-(define_insn "umod<mode>3"
+(define_insn "*umod<mode>3"
[(set (match_operand:GPR 0 "gpc_reg_operand" "=&r")
(umod:GPR (match_operand:GPR 1 "gpc_reg_operand" "r")
(match_operand:GPR 2 "gpc_reg_operand" "r")))]
- "TARGET_MODULO"
+ "TARGET_MODULO && !RS6000_DISABLE_SCALAR_MODULO"
"modu<wd> %0,%1,%2"
[(set_attr "type" "div")
(set_attr "size" "<bits>")])
@@ -3445,7 +3475,7 @@
[(set (match_operand:TI 0 "altivec_register_operand" "=v")
(umod:TI (match_operand:TI 1 "altivec_register_operand" "v")
(match_operand:TI 2 "altivec_register_operand" "v")))]
- "TARGET_POWER10 && TARGET_POWERPC64"
+ "TARGET_POWER10 && TARGET_POWERPC64 && !RS6000_DISABLE_SCALAR_MODULO"
"vmoduq %0,%1,%2"
[(set_attr "type" "vecdiv")
(set_attr "size" "128")])
@@ -3454,7 +3484,7 @@
[(set (match_operand:TI 0 "altivec_register_operand" "=v")
(mod:TI (match_operand:TI 1 "altivec_register_operand" "v")
(match_operand:TI 2 "altivec_register_operand" "v")))]
- "TARGET_POWER10 && TARGET_POWERPC64"
+ "TARGET_POWER10 && TARGET_POWERPC64 && !RS6000_DISABLE_SCALAR_MODULO"
"vmodsq %0,%1,%2"
[(set_attr "type" "vecdiv")
(set_attr "size" "128")])
diff --git a/gcc/testsuite/gcc.target/powerpc/clone1.c b/gcc/testsuite/gcc.target/powerpc/clone1.c
index c69fd2aa1b8..74323ca0e8c 100644
--- a/gcc/testsuite/gcc.target/powerpc/clone1.c
+++ b/gcc/testsuite/gcc.target/powerpc/clone1.c
@@ -21,6 +21,7 @@ long mod_func_or (long a, long b, long c)
return mod_func (a, b) | c;
}
-/* { dg-final { scan-assembler-times {\mdivd\M} 1 } } */
-/* { dg-final { scan-assembler-times {\mmulld\M} 1 } } */
-/* { dg-final { scan-assembler-times {\mmodsd\M} 1 } } */
+/* { Fail due to RS6000_DISABLE_SCALAR_MODULO. */
+/* { dg-final { scan-assembler-times {\mdivd\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmulld\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmodsd\M} 1 { xfail *-*-* } } } */
diff --git a/gcc/testsuite/gcc.target/powerpc/clone3.c b/gcc/testsuite/gcc.target/powerpc/clone3.c
index 911b88b781d..d3eb4dd2378 100644
--- a/gcc/testsuite/gcc.target/powerpc/clone3.c
+++ b/gcc/testsuite/gcc.target/powerpc/clone3.c
@@ -27,7 +27,8 @@ long mod_func_or (long a, long b, long c)
return mod_func (a, b) | c;
}
-/* { dg-final { scan-assembler-times {\mdivd\M} 1 } } */
-/* { dg-final { scan-assembler-times {\mmulld\M} 1 } } */
-/* { dg-final { scan-assembler-times {\mmodsd\M} 2 } } */
+/* { Fail due to RS6000_DISABLE_SCALAR_MODULO. */
+/* { dg-final { scan-assembler-times {\mdivd\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmulld\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmodsd\M} 2 { xfail *-*-* } } } */
/* { dg-final { scan-assembler-times {\mpld\M} 1 } } */
diff --git a/gcc/testsuite/gcc.target/powerpc/mod-1.c b/gcc/testsuite/gcc.target/powerpc/mod-1.c
index 861ba670af4..8720ffb3346 100644
--- a/gcc/testsuite/gcc.target/powerpc/mod-1.c
+++ b/gcc/testsuite/gcc.target/powerpc/mod-1.c
@@ -7,13 +7,14 @@ long lsmod (long a, long b) { return a%b; }
unsigned int iumod (unsigned int a, unsigned int b) { return a%b; }
unsigned long lumod (unsigned long a, unsigned long b) { return a%b; }
-/* { dg-final { scan-assembler-times "modsw " 1 } } */
-/* { dg-final { scan-assembler-times "modsd " 1 } } */
-/* { dg-final { scan-assembler-times "moduw " 1 } } */
-/* { dg-final { scan-assembler-times "modud " 1 } } */
-/* { dg-final { scan-assembler-not "mullw " } } */
-/* { dg-final { scan-assembler-not "mulld " } } */
-/* { dg-final { scan-assembler-not "divw " } } */
-/* { dg-final { scan-assembler-not "divd " } } */
-/* { dg-final { scan-assembler-not "divwu " } } */
-/* { dg-final { scan-assembler-not "divdu " } } */
+/* { Fail due to RS6000_DISABLE_SCALAR_MODULO. */
+/* { dg-final { scan-assembler-times {\mmodsw\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmodsd\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmoduw\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmodud\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mmullw\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mmulld\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mdivw\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mdivd\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mdivwu\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mdivdu\M} { xfail *-*-* } } } */
diff --git a/gcc/testsuite/gcc.target/powerpc/mod-2.c b/gcc/testsuite/gcc.target/powerpc/mod-2.c
index 441ec5878f1..54bdca88607 100644
--- a/gcc/testsuite/gcc.target/powerpc/mod-2.c
+++ b/gcc/testsuite/gcc.target/powerpc/mod-2.c
@@ -5,8 +5,9 @@
int ismod (int a, int b) { return a%b; }
unsigned int iumod (unsigned int a, unsigned int b) { return a%b; }
-/* { dg-final { scan-assembler-times "modsw " 1 } } */
-/* { dg-final { scan-assembler-times "moduw " 1 } } */
-/* { dg-final { scan-assembler-not "mullw " } } */
-/* { dg-final { scan-assembler-not "divw " } } */
-/* { dg-final { scan-assembler-not "divwu " } } */
+/* { Fail due to RS6000_DISABLE_SCALAR_MODULO. */
+/* { dg-final { scan-assembler-times {\mmodsw\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-times {\mmoduw\M} 1 { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mmullw\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mdivw\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler-not {\mdivwu\M} { xfail *-*-* } } } */
diff --git a/gcc/testsuite/gcc.target/powerpc/p10-vdivq-vmodq.c b/gcc/testsuite/gcc.target/powerpc/p10-vdivq-vmodq.c
index 84685e5ff43..148998c8c9d 100644
--- a/gcc/testsuite/gcc.target/powerpc/p10-vdivq-vmodq.c
+++ b/gcc/testsuite/gcc.target/powerpc/p10-vdivq-vmodq.c
@@ -23,5 +23,6 @@ __int128 s_mod(__int128 a, __int128 b)
/* { dg-final { scan-assembler {\mvdivsq\M} } } */
/* { dg-final { scan-assembler {\mvdivuq\M} } } */
-/* { dg-final { scan-assembler {\mvmodsq\M} } } */
-/* { dg-final { scan-assembler {\mvmoduq\M} } } */
+/* { Fail due to RS6000_DISABLE_SCALAR_MODULO. */
+/* { dg-final { scan-assembler {\mvmodsq\M} { xfail *-*-* } } } */
+/* { dg-final { scan-assembler {\mvmoduq\M} { xfail *-*-* } } } */
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2023-10-02 20:06 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-10-02 20:06 [gcc r12-9902] Disable generation of scalar modulo instructions Pat Haugen
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).