From mboxrd@z Thu Jan 1 00:00:00 1970 From: Grant Grundler Subject: Re: [parisc-linux] clerar user page test Date: Thu, 9 Dec 2004 10:32:39 -0700 Message-ID: <20041209173239.GA14560@colo.lackof.org> References: <418A81310000D172@mail-7-bnl.tiscali.it> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: parisc-linux@lists.parisc-linux.org To: Joel Soete Return-Path: In-Reply-To: <418A81310000D172@mail-7-bnl.tiscali.it> List-Id: parisc-linux developers list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: parisc-linux-bounces@lists.parisc-linux.org On Tue, Dec 07, 2004 at 03:12:37PM +0100, Joel Soete wrote: > here are the results of some clup test: ... Joel, I also rewrote the test and parked it on: http://iou.parisc-linux.org/hppa/clup.c Here's the output from a a500-6x running 2.6.10-rc1-pa12-SMP: grundler@iodine:~$ gcc-3.3 -o clup0 clup.c grundler@iodine:~$ gcc-3.3 -DLOOP2X -o clup1 clup.c grundler@iodine:~$ gcc-3.3 -march=2.0 -DLP64 -o clup2 clup.c grundler@iodine:~$ gcc-3.3 -march=2.0 -DLP64 -DLOOP2X -o clup3 clup.c grundler@iodine:~$ gcc-3.3 -march=2.0 -DLP64 -DLOOP2X -DPREFETCH -o clup4 clup.c grundler@iodine:~$ for i in 0 1 2 3 4; do time ./clup$i ; done 2>&1 | fgrep user user 0m2.393s user 0m2.392s user 0m1.238s user 0m1.224s user 0m1.225s I've varied the prefetch values and nothing seems to be helping. I have to wonder if it's because the kernel VM is interferring with the test. But doubling the stride for 64-bit is consistently ~1.8% improvement. grant _______________________________________________ parisc-linux mailing list parisc-linux@lists.parisc-linux.org http://lists.parisc-linux.org/mailman/listinfo/parisc-linux