public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org> To: gcc-bugs@gcc.gnu.org Subject: [Bug target/56309] -O3 optimizer generates conditional moves instead of compare and branch resulting in almost 2x slower code Date: Fri, 15 Feb 2013 09:33:00 -0000 [thread overview] Message-ID: <bug-56309-4-k43fwmIC10@http.gcc.gnu.org/bugzilla/> (raw) In-Reply-To: <bug-56309-4@http.gcc.gnu.org/bugzilla/> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56309 --- Comment #23 from Richard Biener <rguenth at gcc dot gnu.org> 2013-02-15 09:33:22 UTC --- (In reply to comment #20) > (In reply to comment #12) > > --- by-val-O3.s.orig 2013-02-14 18:06:56.000000000 +0100 > > +++ by-val-O3.s 2013-02-14 18:07:23.000000000 +0100 > > @@ -357,9 +357,8 @@ > > shrq $32, %rdi > > cmpq %r8, %rdx > > cmovbe %r11, %rdi > > - addq $1, %rax > > - cmpq %r8, %rdx > > cmovbe %rdx, %rcx > > + addq $1, %rax > > cmpq %rbp, %rax > > movq %rcx, -8(%rsi,%rax,8) > > jne .L50 > > > > unmodified: Took 14.31 seconds total. > > modified: Took 13.04 seconds total. > > > > So re. comment #9: it's not the problem but it'd be a small improvement. > > FWIW this comes from not eliminating the condition expression in > the conditional moves that ifcvt creates: > > tmp_97 = tmp_93 > 4294967295 ? tmp_95 : tmp_93; > carry_105 = tmp_93 > 4294967295 ? carry_94 : 0; > > I'm surprised this form is allowed at all, I'd expect we only allow > is_gimple_reg() for a COND_EXPR_COND in a RHS context. Yeah, it's on my list (even with partial patches available ...). Note that then vectorizing a COND_EXPR is not different from being able to vectorize a comparison statement. pred_2 = tmp_93 > 4294967295; tmp_97 = pred_2 ? tmp_95 : tmp_93; carry_105 = pred_2 ? carry_94 : 0; this form, both vectorized and not vectorized has issues when doing initial instruction selection during expand (with multiple uses of pred_2 we don't TER it). So it might be that we present combine and other RTL optimizers with initial code they will not be able to handle as well as what we do right now. > Anyway -- separate problem. Indeed.
next prev parent reply other threads:[~2013-02-15 9:33 UTC|newest] Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top 2013-02-13 20:21 [Bug middle-end/56309] New: " arturomdn at gmail dot com 2013-02-13 20:23 ` [Bug middle-end/56309] " arturomdn at gmail dot com 2013-02-13 20:26 ` [Bug target/56309] " pinskia at gcc dot gnu.org 2013-02-13 20:29 ` arturomdn at gmail dot com 2013-02-14 8:16 ` ubizjak at gmail dot com 2013-02-14 8:22 ` ubizjak at gmail dot com 2013-02-14 8:28 ` ubizjak at gmail dot com 2013-02-14 14:49 ` ysrumyan at gmail dot com 2013-02-14 15:53 ` arturomdn at gmail dot com 2013-02-14 16:01 ` arturomdn at gmail dot com 2013-02-14 16:44 ` arturomdn at gmail dot com 2013-02-14 16:59 ` steven at gcc dot gnu.org 2013-02-14 17:10 ` steven at gcc dot gnu.org 2013-02-14 17:26 ` [Bug tree-optimization/56309] " steven at gcc dot gnu.org 2013-02-14 17:31 ` arturomdn at gmail dot com 2013-02-14 17:34 ` ubizjak at gmail dot com 2013-02-14 17:43 ` arturomdn at gmail dot com 2013-02-14 17:50 ` ubizjak at gmail dot com 2013-02-14 18:23 ` ubizjak at gmail dot com 2013-02-14 18:35 ` [Bug target/56309] " ubizjak at gmail dot com 2013-02-14 22:18 ` steven at gcc dot gnu.org 2013-02-15 6:50 ` izamyatin at gmail dot com 2013-02-15 7:15 ` jakub at gcc dot gnu.org 2013-02-15 9:33 ` rguenth at gcc dot gnu.org [this message] 2013-02-15 9:36 ` rguenth at gcc dot gnu.org 2013-02-17 8:35 ` [Bug target/56309] conditional moves instead of compare and branch result " ubizjak at gmail dot com 2013-02-17 8:38 ` ubizjak at gmail dot com 2013-02-17 10:10 ` ubizjak at gmail dot com 2013-09-22 6:37 ` icegood1980 at gmail dot com 2013-10-18 9:12 ` jakub at gcc dot gnu.org 2013-10-18 10:06 ` ubizjak at gmail dot com 2015-08-06 9:35 ` ysrumyan at gmail dot com 2021-09-04 20:24 ` peter at cordes dot ca 2021-09-04 20:26 ` peter at cordes dot ca
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=bug-56309-4-k43fwmIC10@http.gcc.gnu.org/bugzilla/ \ --to=gcc-bugzilla@gcc.gnu.org \ --cc=gcc-bugs@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).