On Fri, Jul 26, 2024 at 04:10:48PM +0800, Haochen Jiang wrote:
> * config/i386/avx512dqintrin.h
> (_mm_mask_fpclass_ss_mask): Correct operand order.
> (_mm_mask_fpclass_sd_mask): Ditto.
> (_mm_reduce_round_sd): Use -1 as mask since it is non-mask.
> (_mm_reduce_round_ss): Ditto.
You haven't mentioned the
(_mm_maskz_reduce_round_ss): Use __builtin_ia32_reducess_mask_round
instead of __builtin_ia32_reducesd_mask_round.
change here.
> * config/i386/avx512vlbwintrin.h
> (_mm256_mask_alignr_epi8): Correct operand usage.
> (_mm_mask_alignr_epi8): Ditto.
> * config/i386/avx512vlintrin.h (_mm_mask_alignr_epi64): Ditto.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/i386/avx512bw-vpalignr-1b.c: New test.
> * gcc.target/i386/avx512dq-vfpclasssd-1b.c: Ditto.
> * gcc.target/i386/avx512dq-vfpclassss-1b.c: Ditto.
> * gcc.target/i386/avx512dq-vreducesd-1b.c: Ditto.
> * gcc.target/i386/avx512dq-vreducess-1b.c: Ditto.
> * gcc.target/i386/avx512vl-valignq-1b.c: Ditto.
I went through all the cases and agree with all the changes.
LGTM with the above ChangeLog nit fixed, for trunk/release branches,
even for 14.2 if committed RSN.
Jakub
