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 D4A45385842D; Thu, 22 Feb 2024 17:11:55 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org D4A45385842D
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1708621915;
	bh=5lc1N7uhZSnxyoNovhCZjErEsoAs98YupsP40reRfjY=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=HYOa2npDxePLvQz4ay78NlHfmYmtzTBr0GXuk7cjJT+dLeG+lvvqZH3rY6uzfMCuE
	 ZbX0qPTXT/I7xe2KV+fTiOd66w8VH0yYoEoVLKTx3WG2FsgItN58iv0dZbgPK0JKgW
	 vdwD8wXI0CGTBVxlJYL9qeWZtno5KqbHPJ+o4Jww=
From: "jsm28 at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/111770] predicated loads inactive lane values
 not modelled
Date: Thu, 22 Feb 2024 17:11:51 +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: 14.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jsm28 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: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-111770-4-iEH1fOAvLI@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-111770-4@http.gcc.gnu.org/bugzilla/>
References: <bug-111770-4@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=3D111770

--- Comment #6 from Joseph S. Myers <jsm28 at gcc dot gnu.org> ---
X + 0. -> X is invalid for FP with signed zero or signaling NaN, and also g=
ets
the wrong quantum exponent for decimal FP unless the zero has the maximum
possible quantum exponent (which is not what you get from all bits of the
representation zero, which is a zero with the minimum possible quantum
exponent, or from converting integer 0 to DFP, which has quantum exponent 0=
).

(If you add -0. (with maximum quantum exponent, in the DFP case) instead of
+0., that does produce X for X not a signaling NaN - except in FE_DOWNWARD
mode. Whereas adding +0. is only correct in FE_DOWNWARD mode, since 0. - 0.=
 has
positive sign in all other modes.)=