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 B6AA33858C01; Wed, 23 Aug 2023 13:39:04 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B6AA33858C01
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1692797944;
	bh=TQFBvaeEk4RecHlC8YOgucBiAvs7atsHrOGq/ZCViGQ=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=Gv4tXoRFlPI/U3S/BsxGPEFftL1ylFmXUd9Uy443jp+Vqxykn+tWJcUXrtmj69B35
	 0upW7tgvUGB4ZLbiccMZb8sw+yetP2MRvh1bXpbueahqwkr01+cGlLrojKLPlF/lCt
	 LuvR7jloA6wxjCOSGyo4xKBnOtOV7C9tsBhN7d+U=
From: "crazylht at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/94866] Failure to optimize pinsrq of 0 with index 1 into
 movq
Date: Wed, 23 Aug 2023 13:39:04 +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: 10.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: crazylht at gmail dot com
X-Bugzilla-Status: ASSIGNED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: ubizjak at gmail dot com
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-94866-4-Qn55B56Q0D@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-94866-4@http.gcc.gnu.org/bugzilla/>
References: <bug-94866-4@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=3D94866

--- Comment #6 from Hongtao.liu <crazylht at gmail dot com> ---
(In reply to Uro=C5=A1 Bizjak from comment #4)
> (In reply to Hongtao.liu from comment #3)
> > in x86 backend expand_vec_perm_1, we always tries vec_merge frist for
> > !one_operand_p, expand_vselect_vconcat is only tried when vec_merge fai=
led
> > which means we'd better to use vec_merge instead of vec_select:vec_conc=
at
> > when available in out backend pattern match.
>=20
> In fact, I tried to convert existing sse2_movq128 patterns to vec_merge, =
but
> the patch regressed:
>=20
> -FAIL: gcc.target/i386/sse2-pr94680-2.c scan-assembler movq
> -FAIL: gcc.target/i386/sse2-pr94680-2.c scan-assembler-not pxor
> -FAIL: gcc.target/i386/sse2-pr94680.c scan-assembler-not pxor
> -FAIL: gcc.target/i386/sse2-pr94680.c scan-assembler-times
> (?n)(?:mov|psrldq).*%xmm[0-9] 12
>=20
> So, the compiler still expects vec_concat/vec_select patterns to be prese=
nt.


v2df foo_v2df (v2df x)
 {
   return __builtin_shuffle (x, (v2df) { 0, 0 }, (v2di) { 0, 2 });
 }

The testcase is not a typical vec_merge case, for vec_merge, the shuffle in=
dex
should be {0, 3}. Here it happened to be a vec_merge because the second vec=
tor
is all zero. And yes for this case, we still need to vec_concat:vec_select
pattern.=