Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2021-01-21 Thread Jakub Jelinek via Gcc-patches
On Wed, Jan 06, 2021 at 11:34:32AM +0800, Hongtao Liu via Gcc-patches wrote: > > >> > > >> Note there's a data dependency between them. insn 7 feeds insn 9. When > > >> there's a data dependency, combiner patterns are usually the better > > >> choice than peepholes. I think you'd be looking to m

Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2021-01-05 Thread Hongtao Liu via Gcc-patches
> >> > >> Note there's a data dependency between them. insn 7 feeds insn 9. When > >> there's a data dependency, combiner patterns are usually the better > >> choice than peepholes. I think you'd be looking to match something > >> likethis (from the . combine dump): > >> Using combiner patterns

Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2020-11-30 Thread Jeff Law via Gcc-patches
On 11/16/20 8:10 PM, Hongtao Liu wrote: > On Tue, Nov 17, 2020 at 8:05 AM Jeff Law wrote: >> >> On 9/2/20 3:34 AM, Hongtao Liu via Gcc-patches wrote: >>> Hi: >>> Add define_peephole2 to eliminate potential redundant conversion >>> from mask to vector. >>> Bootstrap is ok, regression test is

Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2020-11-16 Thread Hongtao Liu via Gcc-patches
On Tue, Nov 17, 2020 at 8:05 AM Jeff Law wrote: > > > On 9/2/20 3:34 AM, Hongtao Liu via Gcc-patches wrote: > > Hi: > > Add define_peephole2 to eliminate potential redundant conversion > > from mask to vector. > > Bootstrap is ok, regression test is ok for i386/x86-64 backend. > > Ok for tru

Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2020-11-16 Thread Jeff Law via Gcc-patches
On 9/2/20 3:34 AM, Hongtao Liu via Gcc-patches wrote: > Hi: > Add define_peephole2 to eliminate potential redundant conversion > from mask to vector. > Bootstrap is ok, regression test is ok for i386/x86-64 backend. > Ok for trunk? > > gcc/ChangeLog: > PR target/96891 > * co

Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2020-09-02 Thread H.J. Lu via Gcc-patches
On Wed, Sep 2, 2020 at 2:33 AM Hongtao Liu via Gcc-patches wrote: > > Hi: > Add define_peephole2 to eliminate potential redundant conversion > from mask to vector. > Bootstrap is ok, regression test is ok for i386/x86-64 backend. > Ok for trunk? > > gcc/ChangeLog: > PR target/96891 >

[PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector

2020-09-02 Thread Hongtao Liu via Gcc-patches
Hi: Add define_peephole2 to eliminate potential redundant conversion from mask to vector. Bootstrap is ok, regression test is ok for i386/x86-64 backend. Ok for trunk? gcc/ChangeLog: PR target/96891 * config/i386/sse.md (VI_128_256): New mode iterator. (define_peephol