Re: [PATCH]Don't use pinsr for struct initialization.

2020-08-17 Thread Uros Bizjak via Gcc-patches
On Tue, Aug 18, 2020 at 4:23 AM Hongtao Liu wrote: > > On Fri, Aug 14, 2020 at 5:57 PM Uros Bizjak wrote: > > > > On Fri, Aug 14, 2020 at 8:03 AM Hongtao Liu wrote: > > > > > > Hi: > > > For struct initialization, when it fits in a TImode, gcc will use > > > pinsr insn which causes poor codege

Re: [PATCH]Don't use pinsr for struct initialization.

2020-08-17 Thread Hongtao Liu via Gcc-patches
On Fri, Aug 14, 2020 at 5:57 PM Uros Bizjak wrote: > > On Fri, Aug 14, 2020 at 8:03 AM Hongtao Liu wrote: > > > > Hi: > > For struct initialization, when it fits in a TImode, gcc will use > > pinsr insn which causes poor codegen described in PR93897 and PR96562. > > You should probably remove T

Re: [PATCH]Don't use pinsr for struct initialization.

2020-08-14 Thread Uros Bizjak via Gcc-patches
On Fri, Aug 14, 2020 at 8:03 AM Hongtao Liu wrote: > > Hi: > For struct initialization, when it fits in a TImode, gcc will use > pinsr insn which causes poor codegen described in PR93897 and PR96562. You should probably remove TImode handling also from ix86_expand_pextr. Uros. > Bootstrap i

[PATCH]Don't use pinsr for struct initialization.

2020-08-13 Thread Hongtao Liu via Gcc-patches
Hi: For struct initialization, when it fits in a TImode, gcc will use pinsr insn which causes poor codegen described in PR93897 and PR96562. Bootstrap is ok, regression test is ok for i386/x86-64 backend. Ok for trunk? ChangeLog gcc/ PR target/96562 PR target/93897 *