public inbox for gcc-cvs@sourceware.org help / color / mirror / Atom feed
From: Jakub Jelinek <jakub@gcc.gnu.org> To: gcc-cvs@gcc.gnu.org Subject: [gcc r12-2837] i386: Allow some V32HImode and V64QImode permutations even without AVX512BW [PR80355] Date: Tue, 10 Aug 2021 10:38:33 +0000 (GMT) [thread overview] Message-ID: <20210810103833.870323858007@sourceware.org> (raw) https://gcc.gnu.org/g:50b5877925ef5ae8e9f913d6d2b5ce0204ebc588 commit r12-2837-g50b5877925ef5ae8e9f913d6d2b5ce0204ebc588 Author: Jakub Jelinek <jakub@redhat.com> Date: Tue Aug 10 12:38:00 2021 +0200 i386: Allow some V32HImode and V64QImode permutations even without AVX512BW [PR80355] When working on the PR, I've noticed we generate terrible code for V32HImode or V64QImode permutations for -mavx512f -mno-avx512bw. Generally we can't do much with such permutations, but since PR68655 we can handle at least some, those expressible using V16SImode or V8DImode permutations, but that wasn't reachable, because ix86_vectorize_vec_perm_const didn't even try, it said without TARGET_AVX512BW it can't do anything, and with it can do everything, no d.testing_p attempts. This patch makes it try it for TARGET_AVX512F && !TARGET_AVX512BW. The first hunk is to avoid ICE, expand_vec_perm_even_odd_1 asserts d->vmode isn't V32HImode because expand_vec_perm_1 for AVX512BW handles already all permutations, but when we let it through without !TARGET_AVX512BW, expand_vec_perm_1 doesn't handle it. If we want, that hunk can be dropped if we implement in expand_vec_perm_even_odd_1 and its helper the even permutation as vpmovdw + vpmovdw + vinserti64x4 and odd permutation as vpsrld $16 + vpsrld $16 + vpmovdw + vpmovdw + vinserti64x4. 2021-08-10 Jakub Jelinek <jakub@redhat.com> PR target/80355 * config/i386/i386-expand.c (expand_vec_perm_even_odd): Return false for V32HImode if !TARGET_AVX512BW. (ix86_vectorize_vec_perm_const) <case E_V32HImode, case E_V64QImode>: If !TARGET_AVX512BW and TARGET_AVX512F and d.testing_p, don't fail early, but actually check the permutation. * gcc.target/i386/avx512f-pr80355-2.c: New test. Diff: --- gcc/config/i386/i386-expand.c | 13 +++++++++---- gcc/testsuite/gcc.target/i386/avx512f-pr80355-2.c | 23 +++++++++++++++++++++++ 2 files changed, 32 insertions(+), 4 deletions(-) diff --git a/gcc/config/i386/i386-expand.c b/gcc/config/i386/i386-expand.c index bd21efa9530..c708b33b786 100644 --- a/gcc/config/i386/i386-expand.c +++ b/gcc/config/i386/i386-expand.c @@ -20337,6 +20337,11 @@ expand_vec_perm_even_odd (struct expand_vec_perm_d *d) if (d->perm[i] != 2 * i + odd) return false; + if (d->vmode == E_V32HImode + && d->testing_p + && !TARGET_AVX512BW) + return false; + return expand_vec_perm_even_odd_1 (d, odd); } @@ -20877,16 +20882,16 @@ ix86_vectorize_vec_perm_const (machine_mode vmode, rtx target, rtx op0, return true; break; case E_V32HImode: - if (!TARGET_AVX512BW) + if (!TARGET_AVX512F) return false; - if (d.testing_p) + if (d.testing_p && TARGET_AVX512BW) /* All implementable with a single vperm[it]2 insn. */ return true; break; case E_V64QImode: - if (!TARGET_AVX512BW) + if (!TARGET_AVX512F) return false; - if (d.testing_p) + if (d.testing_p && TARGET_AVX512BW) /* Implementable with 2 vperm[it]2, 2 vpshufb and 1 or insn. */ return true; break; diff --git a/gcc/testsuite/gcc.target/i386/avx512f-pr80355-2.c b/gcc/testsuite/gcc.target/i386/avx512f-pr80355-2.c new file mode 100644 index 00000000000..c510b2f1769 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/avx512f-pr80355-2.c @@ -0,0 +1,23 @@ +/* PR target/80355 */ +/* { dg-do compile } */ +/* { dg-options "-O2 -mavx512f -mno-avx512vl -mno-avx512dq -mno-avx512bw" } */ +/* { dg-final { scan-assembler-times "\tvshufi(?:32x4|64x2)\t" 2 } } */ + +typedef short V __attribute__((vector_size (64))); +typedef char W __attribute__((vector_size (64))); + +W +f0 (W x) +{ + return __builtin_shuffle (x, (W) { 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, + 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, + 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, + 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 }); +} + +V +f1 (V x) +{ + return __builtin_shuffle (x, (V) { 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, + 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 }); +}
reply other threads:[~2021-08-10 10:38 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20210810103833.870323858007@sourceware.org \ --to=jakub@gcc.gnu.org \ --cc=gcc-cvs@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).