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 015F53858403; Fri,  9 Feb 2024 05:10:15 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 015F53858403
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1707455416;
	bh=G1hjR9waV6HlvMlqI0jbiWo6gn0RkcTSTbtP74qjCe0=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=jmVEpVzzdrBCpnTA9Sf1pPHDJD9ByIJtpgQxqpl+dW+JcKmRB31D5Cz9LpOcNSAj4
	 cCbET313lYX21Fk8oIKCq7fVtGyIJ2U5DABPHchP78JMzvVYYwd4++VsnFI1FKfGgx
	 nutj6pawUoVghlAz/YrCb8XjO2NQq6jYunS4XtmM=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/113827] MrBayes benchmark redundant load
Date: Fri, 09 Feb 2024 05:10:15 +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: 14.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: pinskia at gcc dot gnu.org
X-Bugzilla-Status: WAITING
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: 
Message-ID: <bug-113827-4-mQNfh7R7HY@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-113827-4@http.gcc.gnu.org/bugzilla/>
References: <bug-113827-4@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=3D113827

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Robin Dapp from comment #0)
> A hot block in the MrBayes benchmark (as used in the Phoronix testsuite) =
has
> a redundant scalar load when vectorized.
>=20
> Minimal example, compiled with -march=3Drv64gcv -O3
>=20
> int foo (float **a, float f, int n)
> {
>   for (int i =3D 0; i < n; i++)
>     {
>       a[i][0] /=3D f;
>       a[i][1] /=3D f;
>       a[i][2] /=3D f;
>       a[i][3] /=3D f;
>       a[i] +=3D 4;
>     }
> }

LLVM for aarch64 with the above testcase:
``
.L3:
        ldr     x2, [x0]
        mov     x1, x2
        ldr     q31, [x2]
        fdiv    v31.4s, v31.4s, v0.4s
        str     q31, [x1], 16
        str     x1, [x0], 8 ;;;; HERE
        cmp     x3, x0
        bne     .L3
```

There is a store of x1 there.
I really think you messed up reducing the testcase.=