* [r13-8949 Regression] FAIL: gcc.target/i386/avx512dq-vreducess-1b.c (test for excess errors) on Linux/x86_64
@ 2024-07-29 5:28 haochen.jiang
2024-07-29 6:05 ` Jiang, Haochen
0 siblings, 1 reply; 3+ messages in thread
From: haochen.jiang @ 2024-07-29 5:28 UTC (permalink / raw)
To: haochen.jiang, gcc-regression, gcc-patches
On Linux/x86_64,
bb15c4cf21dbe76df5a225342d1fbe8ecd3c7971 is the first bad commit
commit bb15c4cf21dbe76df5a225342d1fbe8ecd3c7971
Author: Haochen Jiang <haochen.jiang@intel.com>
Date: Thu Jul 25 16:12:20 2024 +0800
i386: Fix AVX512 intrin macro typo
caused
FAIL: gcc.target/i386/avx512dq-vreducesd-1b.c (test for excess errors)
FAIL: gcc.target/i386/avx512dq-vreducess-1b.c (test for excess errors)
with GCC configured with
../../gcc/configure --prefix=/export/users/haochenj/src/gcc-bisect/gcc-13/releases/gcc-13/r13-8949/usr --enable-clocale=gnu --with-system-zlib --with-demangler-in-ld --with-fpmath=sse --enable-languages=c,c++,fortran --enable-cet --without-isl --enable-libmpx x86_64-linux --disable-bootstrap
To reproduce:
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --target_board='unix{-m32}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --target_board='unix{-m32\ -march=cascadelake}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --target_board='unix{-m64}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --target_board='unix{-m64\ -march=cascadelake}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --target_board='unix{-m32}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --target_board='unix{-m32\ -march=cascadelake}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --target_board='unix{-m64}'"
$ cd {build_dir}/gcc && make check RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --target_board='unix{-m64\ -march=cascadelake}'"
(Please do not reply to this email, for question about this report, contact me at haochen dot jiang at intel.com.)
(If you met problems with cascadelake related, disabling AVX512F in command line might save that.)
(However, please make sure that there is no potential problems with AVX512.)
^ permalink raw reply [flat|nested] 3+ messages in thread
* RE: [r13-8949 Regression] FAIL: gcc.target/i386/avx512dq-vreducess-1b.c (test for excess errors) on Linux/x86_64
2024-07-29 5:28 [r13-8949 Regression] FAIL: gcc.target/i386/avx512dq-vreducess-1b.c (test for excess errors) on Linux/x86_64 haochen.jiang
@ 2024-07-29 6:05 ` Jiang, Haochen
2024-07-29 6:38 ` [GCC12/13 PATCH] i386: Use _mm_setzero_ps/d instead of _mm_avx512_setzero_ps/d for GCC13/12 Haochen Jiang
0 siblings, 1 reply; 3+ messages in thread
From: Jiang, Haochen @ 2024-07-29 6:05 UTC (permalink / raw)
To: haochen.jiang, gcc-regression, gcc-patches
Ah... When I solved the huge conflict this morning due to AVX10 refactor for
GCC13/12, I forgot that it is in GCC14 AVX10.1 refactor when I added the
_mm_avx512_setzero_ps/pd. Should use _mm_setzero_ps/pd instead.
Never do something tweak quickly on a sleepy Monday morning.
Thx,
Haochen
> -----Original Message-----
> From: haochen.jiang <haochenj@ecsmtp.sh.intel.com>
> Sent: Monday, July 29, 2024 1:28 PM
> To: Jiang, Haochen <haochen.jiang@intel.com>; gcc-regression@gcc.gnu.org;
> gcc-patches@gcc.gnu.org
> Subject: [r13-8949 Regression] FAIL: gcc.target/i386/avx512dq-vreducess-
> 1b.c (test for excess errors) on Linux/x86_64
>
> On Linux/x86_64,
>
> bb15c4cf21dbe76df5a225342d1fbe8ecd3c7971 is the first bad commit
> commit bb15c4cf21dbe76df5a225342d1fbe8ecd3c7971
> Author: Haochen Jiang <haochen.jiang@intel.com>
> Date: Thu Jul 25 16:12:20 2024 +0800
>
> i386: Fix AVX512 intrin macro typo
>
> caused
>
> FAIL: gcc.target/i386/avx512dq-vreducesd-1b.c (test for excess errors)
> FAIL: gcc.target/i386/avx512dq-vreducess-1b.c (test for excess errors)
>
> with GCC configured with
>
> ../../gcc/configure --prefix=/export/users/haochenj/src/gcc-bisect/gcc-
> 13/releases/gcc-13/r13-8949/usr --enable-clocale=gnu --with-system-zlib --
> with-demangler-in-ld --with-fpmath=sse --enable-languages=c,c++,fortran --
> enable-cet --without-isl --enable-libmpx x86_64-linux --disable-bootstrap
>
> To reproduce:
>
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --
> target_board='unix{-m32}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --
> target_board='unix{-m32\ -march=cascadelake}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --
> target_board='unix{-m64}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducesd-1b.c --
> target_board='unix{-m64\ -march=cascadelake}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --
> target_board='unix{-m32}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --
> target_board='unix{-m32\ -march=cascadelake}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --
> target_board='unix{-m64}'"
> $ cd {build_dir}/gcc && make check
> RUNTESTFLAGS="i386.exp=gcc.target/i386/avx512dq-vreducess-1b.c --
> target_board='unix{-m64\ -march=cascadelake}'"
>
> (Please do not reply to this email, for question about this report, contact me at
> haochen dot jiang at intel.com.) (If you met problems with cascadelake
> related, disabling AVX512F in command line might save that.) (However,
> please make sure that there is no potential problems with AVX512.)
^ permalink raw reply [flat|nested] 3+ messages in thread
* [GCC12/13 PATCH] i386: Use _mm_setzero_ps/d instead of _mm_avx512_setzero_ps/d for GCC13/12
2024-07-29 6:05 ` Jiang, Haochen
@ 2024-07-29 6:38 ` Haochen Jiang
0 siblings, 0 replies; 3+ messages in thread
From: Haochen Jiang @ 2024-07-29 6:38 UTC (permalink / raw)
To: gcc-patches; +Cc: hongtao.liu, ubizjak
Hi all,
In GCC13/12, there is no _mm_avx512_setzero_ps/d since it is introduced
in GCC14.
Fix the backport issue as obvious in:
https://gcc.gnu.org/pipermail/gcc-regression/2024-July/080385.html
Thx,
Haochen
gcc/ChangeLog:
* config/i386/avx512dqintrin.h (_mm_reduce_round_sd): Use
_mm_setzero_pd instead of _mm_avx512_setzero_pd.
(_mm_reduce_round_ss): Use _mm_setzero_ps instead of
_mm_avx512_setzero_ps.
---
gcc/config/i386/avx512dqintrin.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/gcc/config/i386/avx512dqintrin.h b/gcc/config/i386/avx512dqintrin.h
index 4383ff14670..82f4acc7d2e 100644
--- a/gcc/config/i386/avx512dqintrin.h
+++ b/gcc/config/i386/avx512dqintrin.h
@@ -2840,7 +2840,7 @@ _mm512_fpclass_ps_mask (__m512 __A, const int __imm)
#define _mm_reduce_round_sd(A, B, C, R) \
((__m128d) __builtin_ia32_reducesd_mask_round ((__v2df)(__m128d)(A), \
- (__v2df)(__m128d)(B), (int)(C), (__v2df) _mm_avx512_setzero_pd (), \
+ (__v2df)(__m128d)(B), (int)(C), (__v2df) _mm_setzero_pd (), \
(__mmask8)(-1), (int)(R)))
#define _mm_mask_reduce_round_sd(W, U, A, B, C, R) \
@@ -2869,7 +2869,7 @@ _mm512_fpclass_ps_mask (__m512 __A, const int __imm)
#define _mm_reduce_round_ss(A, B, C, R) \
((__m128) __builtin_ia32_reducess_mask_round ((__v4sf)(__m128)(A), \
- (__v4sf)(__m128)(B), (int)(C), (__v4sf) _mm_avx512_setzero_ps (), \
+ (__v4sf)(__m128)(B), (int)(C), (__v4sf) _mm_setzero_ps (), \
(__mmask8)(-1), (int)(R)))
#define _mm_mask_reduce_round_ss(W, U, A, B, C, R) \
--
2.31.1
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2024-07-29 6:40 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-07-29 5:28 [r13-8949 Regression] FAIL: gcc.target/i386/avx512dq-vreducess-1b.c (test for excess errors) on Linux/x86_64 haochen.jiang
2024-07-29 6:05 ` Jiang, Haochen
2024-07-29 6:38 ` [GCC12/13 PATCH] i386: Use _mm_setzero_ps/d instead of _mm_avx512_setzero_ps/d for GCC13/12 Haochen Jiang
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).