From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from list by monty-python.gnu.org with tmda-scanned (Exim 4.34)
	id 1BQN7o-0002kS-Ib
	for qemu-devel@nongnu.org; Wed, 19 May 2004 05:12:32 -0400
Received: from mail by monty-python.gnu.org with spam-scanned (Exim 4.34)
	id 1BQN7E-0002ZB-3j
	for qemu-devel@nongnu.org; Wed, 19 May 2004 05:12:27 -0400
Received: from [62.210.158.46] (helo=teheran.magic.fr)
	by monty-python.gnu.org with esmtp (Exim 4.34) id 1BQN78-0002Wh-2j
	for qemu-devel@nongnu.org; Wed, 19 May 2004 05:11:54 -0400
Received: from [10.0.0.2] (ppp-181.net-555.magic.fr [62.210.255.181])
	by teheran.magic.fr (8.11.6/8.11.2) with ESMTP id i4J9BYf03296
	for <qemu-devel@nongnu.org>; Wed, 19 May 2004 11:11:34 +0200 (CEST)
Subject: Re: [Qemu-devel] Power PC speed?
From: "J. Mayer" <l_indien@magic.fr>
In-Reply-To: <000a01c43d30$5509d9d0$0202a8c0@Shaggy>
References: <000a01c43d30$5509d9d0$0202a8c0@Shaggy>
Content-Type: text/plain
Message-Id: <1084957907.8244.6765.camel@rapid>
Mime-Version: 1.0
Date: Wed, 19 May 2004 11:11:48 +0200
Content-Transfer-Encoding: 7bit
Reply-To: qemu-devel@nongnu.org
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://mail.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://mail.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: qemu-devel@nongnu.org

On Wed, 2004-05-19 at 01:32, Jamie Burns wrote:
> Hi,
>  
> With all the commotion surrounding PearPC and its ability to install
> Mac OS X, I was wondering what the rough speed of QMEU's PowerPC
> system emulation? The PearPC JIT system runs at about 1/40th the speed
> of its host. Anybody have a rough idea how the current QEMU compares?

I made differents tests:
Linux kernel compilation and nbench.
I promise to put detailed benches on my page soon :-)
My conclusion is:
the code I'm working on is equivalent to a 200 Mhz PPC on my amd64 3000+
(2Ghz).
Nbench gives 1/15 ratio, which seems coherent.
The CVS code is about two or three times slower because I do a lot of
context synchronisations that a real CPU won't do and this makes context
switches very slow: this really slows down tests like "make olconfig" on
a Linux kernel: makes does a lot of forks.

-- 
J. Mayer <l_indien@magic.fr>
Never organized