public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2024-04-22 5:13 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2024-04-22 5:13 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:9b0c52b69c85acbceccc5676509cfc0fbb41dd52
commit 9b0c52b69c85acbceccc5676509cfc0fbb41dd52
Author: Alexandre Oliva <oliva@adacore.com>
Date: Sun Apr 21 17:24:18 2024 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index cbbfbb24658..0dab2c08acd 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -46,9 +46,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2024-04-22 6:56 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2024-04-22 6:56 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:06cec5e7261a78413d7ccfeea4c8e85b224d370f
commit 06cec5e7261a78413d7ccfeea4c8e85b224d370f
Author: Alexandre Oliva <oliva@adacore.com>
Date: Sun Apr 21 17:24:18 2024 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index cbbfbb24658..0dab2c08acd 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -46,9 +46,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2024-04-21 20:27 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2024-04-21 20:27 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:106f3cf414881d14a285a18919bfec14d2d1ca23
commit 106f3cf414881d14a285a18919bfec14d2d1ca23
Author: Alexandre Oliva <oliva@adacore.com>
Date: Sun Apr 21 17:24:18 2024 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index cbbfbb24658..0dab2c08acd 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -46,9 +46,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2023-04-06 6:34 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2023-04-06 6:34 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:10eb8a39d172b074290f0a05d0e329f79d271fcc
commit 10eb8a39d172b074290f0a05d0e329f79d271fcc
Author: Alexandre Oliva <oliva@adacore.com>
Date: Wed Apr 5 11:26:40 2023 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2023-03-30 14:06 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2023-03-30 14:06 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:f4d75bcb7bcb8abd988f640ecf1e16254c2817ae
commit f4d75bcb7bcb8abd988f640ecf1e16254c2817ae
Author: Alexandre Oliva <oliva@adacore.com>
Date: Thu Mar 30 05:06:41 2023 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2023-03-24 6:24 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2023-03-24 6:24 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:26753fc17871dd5da07cc370257f23911f9faa94
commit 26753fc17871dd5da07cc370257f23911f9faa94
Author: Alexandre Oliva <oliva@adacore.com>
Date: Thu Mar 23 00:44:24 2023 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2023-03-23 3:47 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2023-03-23 3:47 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:59a252fd2cd4de6f0e7f72d5309c12ff929fc934
commit 59a252fd2cd4de6f0e7f72d5309c12ff929fc934
Author: Alexandre Oliva <oliva@adacore.com>
Date: Thu Mar 23 00:44:24 2023 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2023-03-16 14:21 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2023-03-16 14:21 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:bc4a78ffd3a32a1e23def77fe1d5d9987b4866c9
commit bc4a78ffd3a32a1e23def77fe1d5d9987b4866c9
Author: Alexandre Oliva <oliva@adacore.com>
Date: Thu Mar 18 05:35:22 2021 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2023-03-15 14:05 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2023-03-15 14:05 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:4ce3371fd5d163f5170c5154935753c96dd028fa
commit 4ce3371fd5d163f5170c5154935753c96dd028fa
Author: Alexandre Oliva <oliva@adacore.com>
Date: Thu Mar 18 05:35:22 2021 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2021-03-11 10:21 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2021-03-11 10:21 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:178ae14b8d96214d836091b88e6d0fa18fc59086
commit 178ae14b8d96214d836091b88e6d0fa18fc59086
Author: Alexandre Oliva <oliva@adacore.com>
Date: Tue Mar 9 18:43:02 2021 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2021-03-09 21:47 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2021-03-09 21:47 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:4922a0bff033dbdc01c1cb74a8b402e2e7e686ef
commit 4922a0bff033dbdc01c1cb74a8b402e2e7e686ef
Author: Alexandre Oliva <oliva@adacore.com>
Date: Tue Mar 9 18:43:02 2021 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2021-03-09 21:25 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2021-03-09 21:25 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:0993f0765747f0cad8f3bd8c955eb1d0557027e8
commit 0993f0765747f0cad8f3bd8c955eb1d0557027e8
Author: Alexandre Oliva <oliva@adacore.com>
Date: Mon Mar 8 23:22:51 2021 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
* [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b
@ 2021-03-09 2:26 Alexandre Oliva
0 siblings, 0 replies; 13+ messages in thread
From: Alexandre Oliva @ 2021-03-09 2:26 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:246877415ab3cec9e0faaaa5196ed280608019c4
commit 246877415ab3cec9e0faaaa5196ed280608019c4
Author: Alexandre Oliva <oliva@adacore.com>
Date: Mon Mar 8 23:22:51 2021 -0300
adjust vectorization expectations for ppc costmodel 76b
This test expects vectorization at power8+ because strict alignment is
not required for vectors. For power7, vectorization is not to take
place because it's not deemed profitable: 12 iterations would be
required to make it so.
But for power6 and below, the test's 10 iterations are enough to make
vectorization profitable, but the test doesn't expect this. Assuming
the decision is indeed appropriate, I'm adjusting the expectations.
for gcc/testsuite/ChangeLog
* gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c: Adjust
expectations for cpus below power7.
Diff:
---
gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
index 5da4343198c..93798501228 100644
--- a/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
+++ b/gcc/testsuite/gcc.dg/vect/costmodel/ppc/costmodel-vect-76b.c
@@ -45,9 +45,10 @@ int main (void)
return 0;
}
-/* Peeling to align the store is used. Overhead of peeling is too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { vector_alignment_reachable && {! vect_no_align} } } } } */
-/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { vector_alignment_reachable && {! vect_hw_misalign} } } } } */
+/* Peeling to align the store is used. Overhead of peeling is too high
+ for power7, but acceptable for earlier architectures. */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 0 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_no_align} } } } } } */
+/* { dg-final { scan-tree-dump-times "vectorization not profitable" 1 "vect" { target { has_arch_pwr7 && { vector_alignment_reachable && {! vect_hw_misalign} } } } } } */
/* Versioning to align the store is used. Overhead of versioning is not too high. */
-/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || {! vector_alignment_reachable} } } } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" { target { vect_no_align || { {! vector_alignment_reachable} || {! has_arch_pwr7 } } } } } } */
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2024-04-22 6:56 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-04-22 5:13 [gcc(refs/users/aoliva/heads/testme)] adjust vectorization expectations for ppc costmodel 76b Alexandre Oliva
-- strict thread matches above, loose matches on Subject: below --
2024-04-22 6:56 Alexandre Oliva
2024-04-21 20:27 Alexandre Oliva
2023-04-06 6:34 Alexandre Oliva
2023-03-30 14:06 Alexandre Oliva
2023-03-24 6:24 Alexandre Oliva
2023-03-23 3:47 Alexandre Oliva
2023-03-16 14:21 Alexandre Oliva
2023-03-15 14:05 Alexandre Oliva
2021-03-11 10:21 Alexandre Oliva
2021-03-09 21:47 Alexandre Oliva
2021-03-09 21:25 Alexandre Oliva
2021-03-09 2:26 Alexandre Oliva
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).