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/65660] [5 Regression] 252.eon regression on bdver2 with -Ofast Date: Thu, 02 Apr 2015 11:44:00 -0000 [thread overview] Message-ID: <bug-65660-4-rjpZYzbFG1@http.gcc.gnu.org/bugzilla/> (raw) In-Reply-To: <bug-65660-4@http.gcc.gnu.org/bugzilla/> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65660 --- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> --- With a true fix I get t.c:3:20: note: Cost model analysis: Vector inside of loop cost: 4 Vector prologue cost: 13 Vector epilogue cost: 11 Scalar iteration cost: 4 Scalar outside cost: 0 Vector outside cost: 24 prologue iterations: 2 epilogue iterations: 2 Calculated minimum iters for profitability: 7 t.c:3:20: note: Runtime profitability threshold = 6 t.c:3:20: note: Static estimate profitability threshold = 6 thus we still vectorize the loop for bdver2. This is because of an oddity in its cost model which has 6, /* scalar_stmt_cost. */ 4, /* scalar load_cost. */ 4, /* scalar_store_cost. */ 6, /* vec_stmt_cost. */ 0, /* vec_to_scalar_cost. */ 2, /* scalar_to_vec_cost. */ 4, /* vec_align_load_cost. */ 4, /* vec_unalign_load_cost. */ 4, /* vec_store_cost. */ 2, /* cond_taken_branch_cost. */ 1, /* cond_not_taken_branch_cost. */ and thus the prologue/epilogue is not pessimized enough for the extra branches (which are very cheap compared to the scalar and vector stmt costs). I am still testing the patch to avoid the round-off errors and really account scalar stmts correctly. I suppose the EON regression should be fixed by instead avoiding the peeling for alignment with a better idea on cost.
next prev parent reply other threads:[~2015-04-02 11:44 UTC|newest] Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top 2015-04-02 10:19 [Bug target/65660] New: " rguenth at gcc dot gnu.org 2015-04-02 10:19 ` [Bug target/65660] " rguenth at gcc dot gnu.org 2015-04-02 10:29 ` trippels at gcc dot gnu.org 2015-04-02 10:31 ` trippels at gcc dot gnu.org 2015-04-02 10:58 ` rguenth at gcc dot gnu.org 2015-04-02 11:44 ` rguenth at gcc dot gnu.org [this message] 2015-04-02 12:02 ` rguenth at gcc dot gnu.org 2015-04-03 16:22 ` hubicka at gcc dot gnu.org 2015-04-03 16:28 ` hubicka at gcc dot gnu.org 2015-04-03 18:35 ` rguenther at suse dot de 2015-04-04 10:47 ` rguenth at gcc dot gnu.org 2015-04-05 0:27 ` Jan Hubicka 2015-04-04 11:38 ` rguenther at suse dot de 2015-04-05 0:27 ` hubicka at ucw dot cz 2015-04-07 7:26 ` rguenth at gcc dot gnu.org 2015-04-07 7:46 ` rguenth at gcc dot gnu.org 2015-04-07 9:35 ` rguenth at gcc dot gnu.org 2015-04-07 13:41 ` rguenth at gcc dot gnu.org 2015-04-08 5:53 ` hubicka at gcc dot gnu.org 2015-04-08 5:56 ` hubicka at gcc dot gnu.org 2015-04-08 7:26 ` rguenth at gcc dot gnu.org 2015-04-09 0:09 ` hubicka at gcc dot gnu.org 2015-04-09 7:30 ` rguenth at gcc dot gnu.org 2015-04-13 7:34 ` [Bug target/65660] [5/6 " rguenth at gcc dot gnu.org 2015-04-13 7:35 ` [Bug target/65660] [5 " rguenth at gcc dot gnu.org 2015-04-13 7:37 ` rguenth at gcc dot gnu.org 2015-04-22 12:01 ` jakub at gcc dot gnu.org 2015-07-16 9:13 ` rguenth at gcc dot gnu.org
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-65660-4-rjpZYzbFG1@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).