From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id C8676385842C; Tue, 26 Mar 2024 12:02:56 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org C8676385842C DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1711454576; bh=TMw2T5U2URJUpQ3Qskm37oJlLs+VrIbi/GijhiCsXpM=; h=From:To:Subject:Date:In-Reply-To:References:From; b=IlK+tsgGNiaD88i+ZPo/0Fnv28qarTOTNolNo+JrKzWotUVgxbE9cg+U2MbKYPlOK pSR/keuy7C/AL1OeBLlUtSGFmGiQB81Sdx+g+TPul4PEdK5tT00RquXw4s4+J2V1R3 zLkPk7YUfM943fZwXiOVV8HIsFwhaAz4uFS5paI0= From: "rguenth at gcc dot gnu.org" To: gcc-bugs@gcc.gnu.org Subject: [Bug tree-optimization/114057] [14 Regression] 435.gromacs fails verification with -Ofast -march={znver2,znver4} and PGO after r14-7272-g57f611604e8bab Date: Tue, 26 Mar 2024 12:02:55 +0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: gcc X-Bugzilla-Component: tree-optimization X-Bugzilla-Version: 14.0 X-Bugzilla-Keywords: wrong-code X-Bugzilla-Severity: normal X-Bugzilla-Who: rguenth at gcc dot gnu.org X-Bugzilla-Status: NEW X-Bugzilla-Resolution: X-Bugzilla-Priority: P1 X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org X-Bugzilla-Target-Milestone: 14.0 X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: Message-ID: In-Reply-To: References: Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/ Auto-Submitted: auto-generated MIME-Version: 1.0 List-Id: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D114057 --- Comment #9 from Richard Biener --- Interestingly r14-7272-g57f611604e8bab causes quite some BB vectorization cases to be rejected - I would have expected it to only get us more vectorization? -innerf.f:277:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:281:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:284:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:323:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:327:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:330:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:630:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:634:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:637:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:679:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:683:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:686:72: optimized: basic block part vectorized using 8 byte vecto= rs -innerf.f:1283:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:1287:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:1290:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:1354:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:1358:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:1361:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2004:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2008:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2011:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2082:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2086:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2089:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2365:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2369:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:2372:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3347:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3351:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3354:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3390:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3394:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3397:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3434:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3438:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:3441:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4477:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4481:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4484:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4520:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4524:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4527:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4567:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4571:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:4574:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5702:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5706:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5709:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5745:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5749:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5752:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5815:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5819:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:5822:72: optimized: basic block part vectorized using 8 byte vect= ors +innerf.f:6114:72: optimized: basic block part vectorized using 8 byte vect= ors +innerf.f:6118:72: optimized: basic block part vectorized using 32 byte vec= tors +innerf.f:6127:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7050:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7054:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7057:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7093:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7097:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7100:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7170:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7174:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:7177:72: optimized: basic block part vectorized using 8 byte vect= ors +innerf.f:7479:72: optimized: basic block part vectorized using 8 byte vect= ors +innerf.f:7483:72: optimized: basic block part vectorized using 32 byte vec= tors +innerf.f:7492:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:8280:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:8284:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:8287:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9306:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9310:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9313:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9350:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9354:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9357:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9394:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9398:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:9401:72: optimized: basic block part vectorized using 8 byte vect= ors -innerf.f:10483:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10487:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10490:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10527:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10531:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10534:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10574:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10578:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:10581:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11751:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11755:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11758:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11795:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11799:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11802:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11865:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11869:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:11872:72: optimized: basic block part vectorized using 8 byte vec= tors +innerf.f:12173:72: optimized: basic block part vectorized using 8 byte vec= tors +innerf.f:12177:72: optimized: basic block part vectorized using 32 byte vectors +innerf.f:12186:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13136:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13140:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13143:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13180:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13184:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13187:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13257:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13261:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:13264:72: optimized: basic block part vectorized using 8 byte vec= tors +innerf.f:13574:72: optimized: basic block part vectorized using 8 byte vec= tors +innerf.f:13578:72: optimized: basic block part vectorized using 32 byte vectors +innerf.f:13587:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:14631:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:14635:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:14638:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15896:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15900:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15903:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15953:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15957:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15960:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:15998:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:16002:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:16005:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17325:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17329:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17332:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17382:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17386:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17389:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17429:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17433:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:17436:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19038:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19042:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19045:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19095:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19099:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19102:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19164:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19168:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:19171:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20844:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20848:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20851:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20901:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20905:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20908:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20977:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20981:72: optimized: basic block part vectorized using 8 byte vec= tors -innerf.f:20984:72: optimized: basic block part vectorized using 8 byte vec= tors -vec.h:378:9: optimized: basic block part vectorized using 16 byte vectors -vec.h:379:9: optimized: basic block part vectorized using 16 byte vectors -vec.h:380:9: optimized: basic block part vectorized using 16 byte vectors -clincs.c:65:12: optimized: basic block part vectorized using 8 byte vectors -clincs.c:151:12: optimized: basic block part vectorized using 8 byte vecto= rs -vec.h:537:15: optimized: basic block part vectorized using 32 byte vectors -vec.h:453:15: optimized: basic block part vectorized using 32 byte vectors -coupling.c:188:10: optimized: basic block part vectorized using 8 byte vec= tors -vec.h:426:15: optimized: basic block part vectorized using 32 byte vectors -vec.h:252:8: optimized: basic block part vectorized using 8 byte vectors -update.c:266:16: optimized: basic block part vectorized using 8 byte vecto= rs=