From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <dwz-return-557-listarch-dwz=sourceware.org@sourceware.org>
Received: (qmail 42800 invoked by alias); 17 Dec 2019 22:24:19 -0000
Mailing-List: contact dwz-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Id: <dwz.sourceware.org>
List-Post: <mailto:dwz@sourceware.org>
List-Help: <mailto:dwz-help@sourceware.org>
List-Subscribe: <mailto:dwz-subscribe@sourceware.org>
Sender: dwz-owner@sourceware.org
Received: (qmail 42291 invoked by uid 48); 17 Dec 2019 22:24:03 -0000
From: "vries at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>
To: dwz@sourceware.org
Subject: [Bug default/25276] create_import_tree performance improvements
Date: Tue, 01 Jan 2019 00:00:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: dwz
X-Bugzilla-Component: default
X-Bugzilla-Version: unspecified
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: vries at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: nobody at sourceware dot org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-25276-11298-xB4SjdiVZO@http.sourceware.org/bugzilla/>
In-Reply-To: <bug-25276-11298@http.sourceware.org/bugzilla/>
References: <bug-25276-11298@http.sourceware.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://sourceware.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2019-q4/txt/msg00172.txt.bz2

https://sourceware.org/bugzilla/show_bug.cgi?id=3D25276

--- Comment #7 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #6)
> (In reply to Tom de Vries from comment #5)
> > (In reply to Tom de Vries from comment #0)
> > > which is this loop here:
> > > ...
> > >       for (icu =3D ipu->next; icu; icu =3D icu->next)
> > >         icu->seen =3D false;
> > > ...
> > >=20
> > > It might be possible to improve here by keeping track of this data in=
 another
> > > way, say an array of bool or a bitvector, which could be memset to ze=
ro.
> >=20
> > The array of bool + memset approach reduces create_import_tree phase 3 =
time
> > from 104.19 to 23.00.
>=20
> Keeping track of min_seen and max_seen, and only memsetting the affected
> region gets us 21.77.

And the replacing the bool array with a bitvector with base type unsigned i=
nt,
gets us 11.66.

--=20
You are receiving this mail because:
You are on the CC list for the bug.