[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2016-01-21 Thread ienkovich at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 --- Comment #10 from Ilya Enkovich --- Author: ienkovich Date: Thu Jan 21 16:05:14 2016 New Revision: 232680 URL: https://gcc.gnu.org/viewcvs?rev=232680=gcc=rev Log: gcc/ 2016-01-21 Yuri Rumyantsev PR

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2016-01-12 Thread afomin.mailbox at gmail dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 Alexander Fomin changed: What|Removed |Added Status|RESOLVED|CLOSED --- Comment #9 from Alexander

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2016-01-12 Thread afomin.mailbox at gmail dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 Alexander Fomin changed: What|Removed |Added Status|NEW |RESOLVED Resolution|---

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2016-01-11 Thread ienkovich at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 --- Comment #7 from Ilya Enkovich --- Author: ienkovich Date: Mon Jan 11 12:07:31 2016 New Revision: 232220 URL: https://gcc.gnu.org/viewcvs?rev=232220=gcc=rev Log: gcc/ 2016-01-11 Yuri Rumyantsev PR

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-17 Thread ysrumyan at gmail dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 --- Comment #4 from Yuri Rumyantsev --- You are quite right - the cost model is very poor. We did simple experiment and set up the branch cost to 1 but noticed performance regressions on other benchmarks. when we set it to 2 we did not see any

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-17 Thread ubizjak at gmail dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 Uroš Bizjak changed: What|Removed |Added Target|i686-*-*|x86 Status|UNCONFIRMED

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-17 Thread jgreenhalgh at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 --- Comment #6 from James Greenhalgh --- Rather than a target macro, I think I'd rather implement this as a parameter along the lines of "max-rtl-ifcvt-insns" - i386 backend could then set this to 1 and avoid the issue. I'll try to get to that

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-16 Thread jakub at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 Jakub Jelinek changed: What|Removed |Added CC||jakub at gcc dot gnu.org --- Comment #2

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-16 Thread jgreenhalgh at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 --- Comment #3 from James Greenhalgh --- >From the perspective of the (admittedly deficient) very simple cost model in if-conversion, this is not a bad transformation. We replace a branch with cost "3 instructions" with 2 instructions. What is

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-16 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 Richard Biener changed: What|Removed |Added Target Milestone|--- |6.0

[Bug rtl-optimization/68920] [6 Regression] Undesirable if-conversion for a rarely taken branch

2015-12-15 Thread ubizjak at gmail dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68920 --- Comment #1 from Uroš Bizjak --- Another incarnation of PR 56309 ?