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 C276B3858D20; Mon, 15 Jan 2024 15:22:54 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org C276B3858D20
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1705332174;
	bh=j3qG+F1AE88f7O3Lnt2oAG5ExdhQI2SCwbM+Fpg3QdE=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=WcFUYCPIez59IsrMBqlaCwxJUmx7gY6pCvXovZeyt/LekKredI9+pXI1D0rBY3c/0
	 1o17fhbByH1cXUg25ARvE4kY6KpNvn+qxXsdqfY9iWNbRvgZkFJIf6CmsTA/1Iemyy
	 8xTzttAnX8pwQIBQ246isYPjD1/EwAy0q6qGvnn0=
From: "newbie-02 at gmx dot de" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/113365] LONG DOUBLE: denormals: assigning a constant:
 factor 100 slow,
Date: Mon, 15 Jan 2024 15:22:53 +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: unknown
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: newbie-02 at gmx dot de
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Resolution: INVALID
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-113365-4-mJIpF7kSja@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-113365-4@http.gcc.gnu.org/bugzilla/>
References: <bug-113365-4@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=3D113365

--- Comment #4 from newbie-02 <newbie-02 at gmx dot de> ---
hello @Andrew Pinski,=20=20=20

just if I'm allowed to add one more point / question:=20=20

using the testing program, playing with optimization, I get=20=20
the following results for binary64s ( doubles ):=20=20

unoptimized:=20=20
time:     result:                    calculation:=20=20
0.005571; 4.940656458412465442E-324; x2d =3D nextafter( 1E-323, 0.0 )=20
0.002029; 4.940656458412465442E-324; x2d =3D ( ( 1E-323 ) - 5E-324 )=20
0.005395; 1.482196937523739633E-323; x2d =3D nextafter( 1E-323, 1.0 )=20
0.002023; 1.482196937523739633E-323; x2d =3D ( ( 1E-323 ) + 5E-324 )=20
which points out that addition / subtraction are faster in denormals,=20=20
or assignd at compile time despite x2d is defined volatile.=20=20

optimized ( compiled with -O2 ): I mostly! get understandable results,=20=20
( nextafter similar and ADD / SUB faster reg, compiler cheating?,=20=20
but on some occasions ( ~10% of compile attempts ):=20=20
time:     result:                    calculation:=20=20
0.009311; 4.940656458412465442E-324; x2d =3D nextafter( 1E-323, 0.0 )=20
0.000420; 4.940656458412465442E-324; x2d =3D ( ( 1E-323 ) - 5E-324 )=20
0.009668; 1.482196937523739633E-323; x2d =3D nextafter( 1E-323, 1.0 )=20
0.000423; 1.482196937523739633E-323; x2d =3D ( ( 1E-323 ) + 5E-324 )=20
which has ** ~doubled execution time for nextafters **.=20=20

Is that something I should worry about? Somthing which could be improved?=
=20=20
Either here by options or in gcc?=