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 F0B90385841C; Sun, 24 Sep 2023 16:33:21 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org F0B90385841C
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1695573201;
	bh=1aVx9XVzDSMx3sqShoAppgN4s9iibAX7megSw3cEnu4=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=Ibl1VXGRlcTcvzJ0X0kHBL3Hem+pprBGBn86pRovB9Z3uR7n3tfgCTi/eitwezRDk
	 AsqRi+51pJZLqOjvO1GwDBiDBTPnwU19E7GhWNK4uskLe7dHnsg1+g3H/t4+6tEOhP
	 tjPyr35QFkn5sRTmFmQrF+lPkoRUnMzN4FCxxR44=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/111563] Missed optimization of LICM
Date: Sun, 24 Sep 2023 16:33:21 +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: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pinskia at gcc dot gnu.org
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-111563-4-xjgPicbQPU@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-111563-4@http.gcc.gnu.org/bugzilla/>
References: <bug-111563-4@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=3D111563

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Yi from comment #2)
> (In reply to Andrew Pinski from comment #1)
> > _5 =3D var_0_16(D) + var_6_18(D);
> >   invariant up to level 1, cost 1.
> >=20
> > Basically because the cost is not high enough ...
> >=20
> > If you use   --param=3Dlim-expensive=3D1. then it will pull it out of t=
he loop.
> >=20
> > So the cost model is doing the correct thing here ...
>=20
>=20
> Thank you very much for your prompt reply! It took me some time to confirm
> our work. For Example 1, GCC does exactly what you say it does. But for
> Example 2, it doesn't seem to work as expected.
>=20
> We note that `(-(200 / var_10)) + (-var_8) + var_14 + var_2` as a whole c=
an
> be treated as a loop invariant, but gcc-trunk -O3 --param=3Dlim-expensive=
=3D1
> does not:

  _4 =3D _2 + var_24_lsm.5_21;
  _6 =3D _4 + _5;

So this is again reassociation with LIM, the same issue as PR 111560.=