From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id 73B6E385841A; Wed, 1 Feb 2023 08:19:44 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 73B6E385841A DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1675239584; bh=dmTaZVp3CWy2Xao4L+c45XUrfvEamsEcq+zxuBBUcrY=; h=From:To:Subject:Date:In-Reply-To:References:From; b=hC6ESdsAlZoQ7zbJp6HDgFBX0cdkVUsNdcj8gMAnSfTgyQSwIaapbW+kPpENNHpld aCbDfO5x7beyhT9zW4eqXFULk5x2Eq18CDjwyWUYN9iZ+CYatqWPRVvR8zK3TEJjhV aidk4eHbNDIklgm1ObNYMh0WcvYgs1LxxWbMiYP8= From: "manolis.tsamis at vrull dot eu" To: gcc-bugs@gcc.gnu.org Subject: [Bug tree-optimization/98138] BB vect fail to SLP one case Date: Wed, 01 Feb 2023 08:19:42 +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: 11.0 X-Bugzilla-Keywords: missed-optimization X-Bugzilla-Severity: normal X-Bugzilla-Who: manolis.tsamis at vrull dot eu X-Bugzilla-Status: NEW X-Bugzilla-Resolution: X-Bugzilla-Priority: P3 X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org X-Bugzilla-Target-Milestone: --- X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: cc 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=3D98138 manolis.tsamis at vrull dot eu changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |manolis.tsamis at vrull do= t eu --- Comment #11 from manolis.tsamis at vrull dot eu --- > The full satd_8x4 looks like the following, the 2nd loop isn't to be > disregarded Regarding the second loop, this patch https://gcc.gnu.org/pipermail/gcc-patches/2022-December/608827.html should result in improved vectorization and performance. Currently ((a>>15)&0x10001)*0xffff from abs2 is calculated using 4 vector operations (shift, bitand, shift+sub for the multiplication) whereas with t= he proposed patch this will be a single vector compare operation.=