[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 Martin Liška changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #13 from Martin Liška --- Closing.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 --- Comment #12 from Martin Liška --- (In reply to amker from comment #11) > (In reply to Martin Liška from comment #10) > > Let me bisect that. > > IIRC, it's one of the IVOPTs rewriting patches. The Bisected revision may > not be appropriate for backport on itself I think. Ok, so then probably close the PR as we're not planning a rewritten to gcc7 branch?
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 --- Comment #11 from amker at gcc dot gnu.org --- (In reply to Martin Liška from comment #10) > Let me bisect that. IIRC, it's one of the IVOPTs rewriting patches. The Bisected revision may not be appropriate for backport on itself I think.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 Martin Liška changed: What|Removed |Added Keywords||needs-bisection Assignee|unassigned at gcc dot gnu.org |marxin at gcc dot gnu.org --- Comment #10 from Martin Liška --- Let me bisect that.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 --- Comment #9 from Jeffrey A. Law --- Leslie, you'd need to bisect. Probably something from Bin in the summer of 2017. Not something we're likely to backport.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 Leslie Zhai changed: What|Removed |Added CC||lesliezhai at llvm dot org.cn --- Comment #8 from Leslie Zhai --- (In reply to Jeffrey A. Law from comment #7) > Fixed on the trunk. Hi Jeff, Could you please show me which commit fixed it on trunk? thanks a lot! Regards, Leslie Zhai
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 Jeffrey A. Law changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2018-01-08 CC||law at redhat dot com Summary|[7/8 Regression] Changes in |[7 Regression] Changes in |ivopts caused perf |ivopts caused perf |regression on x86 |regression on x86 Ever confirmed|0 |1 --- Comment #7 from Jeffrey A. Law --- Fixed on the trunk.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 --- Comment #4 from Jakub Jelinek --- So shall we defer this PR to GCC 8 then (i.e. [8 Regression] and Target Milestone: 8.0? Richard, are you ok with that?
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 --- Comment #3 from amker at gcc dot gnu.org --- (In reply to Jakub Jelinek from comment #2) > Any progress with this? I had a patch which can resolve the reported regression, and generate even better code than the original: .L5: movl(%esi,%ecx,4), %eax movl-4128(%ebp), %edx movl-4124(%ebp), %ebx imull (%edi,%ecx,4), %ebx imull %eax, %edx imull -4124(%ebp), %eax addl%ebx, %edx movl-4128(%ebp), %ebx imull (%edi,%ecx,4), %ebx addl%ebx, %eax movl-4132(%ebp), %ebx movl%edx, (%ebx,%ecx,4) movl-4136(%ebp), %ebx movl%edx, (%edi,%ecx,4) movl%eax, (%ebx,%ecx,4) movl%eax, (%esi,%ecx,4) addl-4140(%ebp), %ecx cmpl$511, %ecx jle .L5 But this patch is a complete rewrite of major part of IVOPT, which is a stage 1 work.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 Jakub Jelinek changed: What|Removed |Added CC||jakub at gcc dot gnu.org --- Comment #2 from Jakub Jelinek --- Any progress with this?
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 amker at gcc dot gnu.org changed: What|Removed |Added CC||amker at gcc dot gnu.org --- Comment #1 from amker at gcc dot gnu.org --- It doesn't look like a regression from the dump, I suspect it's because how gcc handles symbol (arr_1/arr_2) in m32 PIE code. I will have a look. BTW, the patch itself is right, it triggers cost model issue again in which wrong/inaccurate cost gives better result. I am doing experiments rewriting the whole cost computation part.
[Bug tree-optimization/71361] [7 Regression] Changes in ivopts caused perf regression on x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71361 Richard Biener changed: What|Removed |Added Keywords||missed-optimization Target Milestone|--- |7.0