public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
From: "eric.batut at allegorithmic dot com" <gcc-bugzilla@gcc.gnu.org> To: gcc-bugs@gcc.gnu.org Subject: [Bug target/54300] New: [4.7/4.8 Regression] Erroneous optimization causes wrong Neon data management Date: Fri, 17 Aug 2012 15:28:00 -0000 [thread overview] Message-ID: <bug-54300-4@http.gcc.gnu.org/bugzilla/> (raw) http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54300 Bug #: 54300 Summary: [4.7/4.8 Regression] Erroneous optimization causes wrong Neon data management Classification: Unclassified Product: gcc Version: 4.8.0 Status: UNCONFIRMED Severity: major Priority: P3 Component: target AssignedTo: unassigned@gcc.gnu.org ReportedBy: eric.batut@allegorithmic.com Created attachment 28044 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28044 Small repro case Using gcc trunk at rev 190381, compiled with the Android NDK r8b build-gcc.sh script (so an arm-linux-androideabi target) and the command line below, the attached repro case generates wrong code: arm-linux-androideabi-g++ -march=armv7-a -mfloat-abi=softfp -mfpu=vfp -mfpu=neon -marm -O2 test.cpp -S -o test.s The core loop is pasted below: for(unsigned int sv=0 ; sv!=dv0 ; sv=(sv+s1v)&smask_v) { int32x4_t s; s = vmovl_s16(vget_low_s16(_loadlo_8i16((cv8i16*)_Inp, sv ))); c = vaddq_s32(c, s); } 8 bytes are fetched from "_Inp (in bytes) + sv", then sign-extended from 4 16bits values to 4 32bits values, then accumulated in "c". The generated assembly code for the loop is: .L3: add r4, r0, ip vmov.i32 d18, #0 @ v4hi <= d18 is full of 0's add ip, ip, r2 vld1.16 {d19}, [r4:64] <= d19 holds useful data and ip, ip, r5 cmp r3, ip vswp d18, d19 <= d19 is now full of 0's vmovl.s16 q9, d19 <= d19 (full of 0's) gets expanded vadd.i32 q8, q8, q9 <= q9 is always zero when accumulated bne .L3 When using "-O1" or "-O2 -fno-gcse", correct code is generated: .L3: add r4, r0, ip add ip, ip, r2 and ip, ip, r5 vld1.16 {d18}, [r4:64] <= d18 holds useful data cmp r3, ip vmovl.s16 q9, d18 <= d18 is sign-extended vadd.i32 q8, q8, q9 <= q9 is accumulated bne .L3 This also happens with gcc 4.7.1, but not with gcc 4.6 Also, in the loadlo_8i16 function, if we replace the call to zero_64 by the proper vdup_n_s16(0), then correct code is generated at -O2. The (stripped down in the repro case) _v16u8_ and _v8u8_ structures are the way we implemented some kind of compiler-performed polymorphism for Neon variables, since not all ARM compilers have -flax-vector-conversions.
next reply other threads:[~2012-08-17 15:28 UTC|newest] Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top 2012-08-17 15:28 eric.batut at allegorithmic dot com [this message] 2012-08-18 8:12 ` [Bug target/54300] " rguenth at gcc dot gnu.org 2012-08-20 16:11 ` eric.batut at allegorithmic dot com 2012-09-07 10:33 ` rguenth at gcc dot gnu.org 2012-09-20 10:26 ` jakub at gcc dot gnu.org 2012-10-25 12:56 ` eric.batut at allegorithmic dot com 2012-10-25 23:03 ` steven at gcc dot gnu.org 2012-12-03 15:42 ` rguenth at gcc dot gnu.org 2013-01-10 22:19 ` [Bug target/54300] [4.7 " pinskia at gcc dot gnu.org 2013-01-11 10:42 ` eric.batut at allegorithmic dot com 2013-04-11 8:00 ` rguenth at gcc dot gnu.org 2013-08-20 16:32 ` eric.batut at allegorithmic dot com 2013-10-04 17:29 ` [Bug rtl-optimization/54300] " rearnsha at gcc dot gnu.org 2013-10-08 9:36 ` [Bug rtl-optimization/54300] [4.7, 4.8, 4.9 Regression] regcprop incorrectly looks through parallel register swap operation ebotcazou at gcc dot gnu.org 2013-10-08 9:47 ` rearnsha at gcc dot gnu.org 2013-10-08 10:54 ` ebotcazou at gcc dot gnu.org 2013-11-20 13:55 ` rearnsha at gcc dot gnu.org 2013-12-05 9:56 ` [Bug rtl-optimization/54300] [4.7, 4.8 " steven at gcc dot gnu.org 2013-12-09 14:54 ` rearnsha at gcc dot gnu.org 2014-01-09 15:19 ` rearnsha at gcc dot gnu.org 2014-01-10 16:55 ` rearnsha at gcc dot gnu.org 2014-01-10 16:56 ` rearnsha at gcc dot gnu.org 2014-08-29 1:27 ` gregory.0xf0 at gmail dot com
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-54300-4@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).