From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:47824) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1UaEin-0007YY-E0 for qemu-devel@nongnu.org; Wed, 08 May 2013 20:28:34 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1UaEim-00050L-21 for qemu-devel@nongnu.org; Wed, 08 May 2013 20:28:33 -0400 Received: from mail-ie0-x22b.google.com ([2607:f8b0:4001:c03::22b]:35683) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1UaEil-000503-UI for qemu-devel@nongnu.org; Wed, 08 May 2013 20:28:32 -0400 Received: by mail-ie0-f171.google.com with SMTP id e11so4459440iej.30 for ; Wed, 08 May 2013 17:28:30 -0700 (PDT) Date: Wed, 08 May 2013 19:28:27 -0500 From: Rob Landley References: <8661yvqasu.fsf@shell.gmplib.org> <20130507225736.GO5000@ohm.aurel32.net> <5189FDD3.5000902@redhat.com> <86fvxxoldi.fsf@shell.gmplib.org> In-Reply-To: <86fvxxoldi.fsf@shell.gmplib.org> (from tg@gmplib.org on Wed May 8 04:45:45 2013) Message-Id: <1368059307.18069.219@driftwood> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; DelSp=Yes; Format=Flowed Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] Reporting Heisenbugs in qemu List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Torbjorn Granlund Cc: Paolo Bonzini , Aurelien Jarno , Artyom Tarasenko , qemu-devel On 05/08/2013 04:45:45 AM, Torbjorn Granlund wrote: > Paolo Bonzini writes: >=20 > I guess that's the register windows. There's only so much you can =20 > do to > optimize them, and heavily recursive workloads (like Perl, or the =20 > RTL > half of GCC) pay a hefty price. >=20 > Two qemu targets stand out for slowness, sparc (32 and 64) and mips =20 > (64, > don't know about 32). >=20 > x86 (32 and 64), arm, and ppc run with a slowdown of < 30 for my bogus > benchmark of GMP configure+make. >=20 > With FreeBSD x86_64 I see a slowdown of just 13. (My reference system > runs FreeBSD, so running FreeBSD under qemu is only far.) >=20 > My claimed slowdown factors are affected by kernel, libraries, and > unfortunately very much by gcc speed, which vary with target. >=20 > If the sparc emulation speed is due to register windows, then why does > mips seem just as slow? >=20 > If register windows shortage is a problem, it should be easy to =20 > pretend > to have lots of them, right? sh4 is pretty slow too. Unfortunately: http://landley.net/aboriginal/bin/system-image-sh4.tar.bz2 Only has 64 megs of memory in the emulated board. (Enough to build =20 hello world, not enough to build most packages.) I have a vague todo =20 item to add a command line thing to qemu to plug a physical memory =20 address range into an aribtrary address and then tell linux =20 discontigmem "add memory range HERE" on the command line. That way I =20 wouldn't have to hack up each board emulation to get more memory...) Rob=