From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id B4F093858D35; Mon,  7 Nov 2022 06:58:23 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B4F093858D35
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1667804303;
	bh=9XprcNods5mihx8t+aaxGuYs5xfkT3HgG4I0ltQaA2I=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=bh9I+ffiQEc3yxAKd8nssU0lHPBJG6vfgpe9L0T0w5JsqCXI/l+Y3KBmBs+yUvtyU
	 OfDn9zL5taWZoSqGeZ1wm6ptBPxmtJDJWsQkMd2cxukWufyDa5IdI5l00H5t8BgZfK
	 wejLYiX0bvZvzAX8sagqqjBmW+clEwiJFTFxr1a8=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/107546] [10/11/12/13 Regression] simd, redundant pcmpeqb
 and pxor
Date: Mon, 07 Nov 2022 06:58:23 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 12.2.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pinskia at gcc dot gnu.org
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: 10.5
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: short_desc target_milestone
Message-ID: <bug-107546-4-GZ09FuIoni@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-107546-4@http.gcc.gnu.org/bugzilla/>
References: <bug-107546-4@http.gcc.gnu.org/bugzilla/>
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: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D107546

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|simd, redundant pcmpeqb and |[10/11/12/13 Regression]
                   |pxor                        |simd, redundant pcmpeqb and
                   |                            |pxor
   Target Milestone|---                         |10.5

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
For GNU C++ vectors produced
GCC 4.8 until GCC 11 produced:
        movdqa  xmm0, XMMWORD PTR [rdi]
        pcmpeqd xmm1, xmm1
        pcmpgtb xmm0, XMMWORD PTR .LC0[rip]
        pandn   xmm0, xmm1
        ret

GCC 11+ produces:
        movdqa  xmm0, XMMWORD PTR [rdi]
        pxor    xmm1, xmm1
        pcmpgtb xmm0, XMMWORD PTR .LC0[rip]
        pcmpeqb xmm0, xmm1
        ret

But the intrinics produced the expected thing until GCC 9.

in GCC 8 the intrinsics produces:
  _3 =3D VEC_COND_EXPR <_4 < { 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 48, =
48,
48, 48, 48, 48 }, { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,=
 -1,
-1 }, { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }>;

even. Notice the < vs <=3D there.
I suspect the <=3D expansion part of the x86_64 backend needs to be fixed u=
p to
produce better code.

So this is a regression for the intrinsics and marking it as such.=