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 1652E3857C70; Sat,  7 Nov 2020 21:22:16 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 1652E3857C70
From: "wilson at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/94083] inefficient soft-float x!=Inf code
Date: Sat, 07 Nov 2020 21:22:16 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: middle-end
X-Bugzilla-Version: 9.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: minor
X-Bugzilla-Who: wilson at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
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-94083-4-GGpKiMjrtQ@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-94083-4@http.gcc.gnu.org/bugzilla/>
References: <bug-94083-4@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
X-BeenThere: gcc-bugs@gcc.gnu.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Gcc-bugs mailing list <gcc-bugs.gcc.gnu.org>
List-Unsubscribe: <https://gcc.gnu.org/mailman/options/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=unsubscribe>
List-Archive: <https://gcc.gnu.org/pipermail/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-request@gcc.gnu.org?subject=help>
List-Subscribe: <https://gcc.gnu.org/mailman/listinfo/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=subscribe>
X-List-Received-Date: Sat, 07 Nov 2020 21:22:16 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D94083

The original bug report was apparently lost in the sourceware/gcc migration
back in the spring and I didn't notice until now.

This testcase

int foo(void) {
  volatile float f, g;
  int            n;
  f =3D __builtin_huge_valf();
  g =3D __builtin_huge_valf();
  n +=3D 1 - (f !=3D __builtin_huge_valf());
  return n;
}

compiled for soft-float with -O2, and looking at the original tree dump I s=
ee

  f =3D  Inf;
  g =3D  Inf;
  SAVE_EXPR <!(f u<=3D 3.4028234663852885981170418348451692544e+38)>;, n =3D
SAVE_EX
PR <!(f u<=3D 3.4028234663852885981170418348451692544e+38)> + n;;

So the C front end converted the f !=3D Inf compare to a f u<=3D
<max-representable-float> compare, but the problem here is that the !=3D
operation is a single libcall, but u<=3D is two libcalls.  So code that sho=
uld
have a single soft-float libcall ends up with two.  First a call to __unord=
sf2,
then a compare and branch, and then a call to __lesf2.  This is a
de-optimization.

Perhaps we can convert the f u<=3D <max-representable-float> back to f !=3D=
 Inf in
the optimization to get a single libcall.  Or maybe we can add unordered
soft-float libcalls like ulesf2.=