public inbox for gcc-cvs@sourceware.org help / color / mirror / Atom feed
From: Richard Biener <rguenth@gcc.gnu.org> To: gcc-cvs@gcc.gnu.org Subject: [gcc(refs/users/rguenth/heads/vect-force-slp)] Allow single-lane SLP of not pattern detected gather/scatter Date: Mon, 16 Oct 2023 12:49:51 +0000 (GMT) [thread overview] Message-ID: <20231016124951.1BC9E3857709@sourceware.org> (raw) https://gcc.gnu.org/g:a01f778e3e051f4b87f167a71abcc1fef878c7ef commit a01f778e3e051f4b87f167a71abcc1fef878c7ef Author: Richard Biener <rguenther@suse.de> Date: Fri Sep 29 13:02:23 2023 +0200 Allow single-lane SLP of not pattern detected gather/scatter The following allows to form single-lane SLP graphs from gathers and scatters that were not turned into the internal function representation by pattern matching. That should help both x86 and emulated gather/scatter to survive single-lane SLP build. * tree-vect-slp.cc (vect_build_slp_tree_1): Allow single-lane gather/scatter. Diff: --- gcc/tree-vect-slp.cc | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/gcc/tree-vect-slp.cc b/gcc/tree-vect-slp.cc index ef3a105cfaf4..97bd2f3feea2 100644 --- a/gcc/tree-vect-slp.cc +++ b/gcc/tree-vect-slp.cc @@ -1360,7 +1360,7 @@ vect_build_slp_tree_1 (vec_info *vinfo, unsigned char *swap, splats the same we handle single element interleaving. */ && (is_a <bb_vec_info> (vinfo) || stmt_info != first_stmt_info - || STMT_VINFO_GATHER_SCATTER_P (stmt_info))) + || (STMT_VINFO_GATHER_SCATTER_P (stmt_info) && i != 0))) { /* Not grouped load. */ if (dump_enabled_p ()) @@ -1861,6 +1861,8 @@ vect_build_slp_tree_2 (vec_info *vinfo, slp_tree node, gcc_assert (gimple_call_internal_p (stmt, IFN_MASK_LOAD) || gimple_call_internal_p (stmt, IFN_GATHER_LOAD) || gimple_call_internal_p (stmt, IFN_MASK_GATHER_LOAD)); + else if (STMT_VINFO_GATHER_SCATTER_P (stmt_info)) + ; else { *max_nunits = this_max_nunits;
next reply other threads:[~2023-10-16 12:49 UTC|newest] Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top 2023-10-16 12:49 Richard Biener [this message] -- strict thread matches above, loose matches on Subject: below -- 2023-10-06 7:07 Richard Biener
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=20231016124951.1BC9E3857709@sourceware.org \ --to=rguenth@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).