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 03CCC3858C60; Thu,  9 Feb 2023 08:47:41 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 03CCC3858C60
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1675932462;
	bh=ixDIEWn3unh+m1G9w46jrpkNBk2fQw9ONoLuRpQ1M/0=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=oPONRtMxRzYlyOBf1a8ErFniWIkujJeWw6Sm97LXBIMpXXuNupH6Fckra7DkY30U/
	 BQ+11dStf/cY8IaudE2NAsI3avieN1GOYrkCAPPjs4HxAnP+s+tmIttuahPPqZqdG+
	 tT0r7PY//Tfr2NTDfAr65RakJ/CZkj+7b5AmGMYI=
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/26854] Inordinate compile times on large
 routines
Date: Thu, 09 Feb 2023 08:47:39 +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: 4.3.0
X-Bugzilla-Keywords: compile-time-hog, memory-hog
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenth at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-26854-4-6f2eNmlZXJ@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-26854-4@http.gcc.gnu.org/bugzilla/>
References: <bug-26854-4@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=3D26854

--- Comment #149 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #148)
> (In reply to lucier from comment #145)
> > Created attachment 54424 [details]
> > CPU and Memory usage reports for mainline 13.0.1 (mainline)
> >=20
> > Thank you for looking at this issue again.
> >=20
> > I built today's mainline and ran three tests:
> >=20
> > /pkgs/gcc-mainline/bin/gcc -v -c -O2 -fmem-report -ftime-report compile=
r.i
> > -save-temps >& report-compiler4
> > /pkgs/gcc-mainline/bin/gcc -v -c -O2 -fmem-report -ftime-report all.i
> > -save-temps >& report-all4
> > /pkgs/gcc-mainline/bin/gcc -v -c -O2 -fmem-report -ftime-report _num.i
> > -save-temps >& report-_num4
> >=20
> > The reports are collected here.  Compiling compiler.i required over 30G=
B at
> > certain points.  I haven't really studied these types of reports in a w=
hile,
> > so I can offer little analysis.
>=20
> Yes, I've mostly looked at performance with -O1 since that's what we sugg=
est
> if -O2+ run into issues.  That seems to be in check with GCC 13, memory
> peak is at ~1GB for me for compiler.i
>=20
> I'll look at -O2 now, the mdreorg thing seems "new" (DF RD is a known sou=
rce
> of issues here ...).

So with -O2 -mno-stv we have

 df reaching defs                   :  18.54 (  9%)
 df use-def / def-use chains        :  10.37 (  5%)
 parser function body               :  20.29 ( 10%)
 tree VRP                           :  17.82 (  9%)
 tree PTA                           :  45.85 ( 23%)
 TOTAL                              : 202.90

and a peak memory use of still 26GB.=