From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.sysgo.com (mail.sysgo.com [62.8.134.5]) by ozlabs.org (Postfix) with ESMTP id 1AE4D679EB for ; Tue, 14 Mar 2006 22:56:42 +1100 (EST) Received: from localhost (localhost [127.0.0.1]) by mail.sysgo.com (Postfix) with ESMTP id E801BFB881 for ; Tue, 14 Mar 2006 12:56:33 +0100 (CET) Received: from donald.sysgo.com (unknown [172.20.1.30]) by mail.sysgo.com (Postfix) with ESMTP id CAE11FB881 for ; Tue, 14 Mar 2006 12:56:33 +0100 (CET) Received: from cam (unknown [172.40.1.200]) by donald.sysgo.com (Postfix) with ESMTP id F02612016DC for ; Tue, 14 Mar 2006 12:56:32 +0100 (CET) From: Carlos Mitidieri To: linuxppc-embedded@ozlabs.org Subject: Re: Stable Linux kernel 2.6 for MPC8XX Date: Tue, 14 Mar 2006 12:56:24 +0100 References: <20060310153341.080163525CB@atlas.denx.de> <200603140850.47014.david.jander@protonic.nl> In-Reply-To: <200603140850.47014.david.jander@protonic.nl> MIME-Version: 1.0 Content-Type: Multipart/Mixed; boundary="Boundary-00=_p9qFEshaJsSESxM" Message-Id: <200603141256.25652.carlos.mitidieri@sysgo.com> List-Id: Linux on Embedded PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , --Boundary-00=_p9qFEshaJsSESxM Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Hi, It seems that the nbench is not the most appropriate suit for comparing OSs (or OS versions), since its programs exercise only the CPU/memory subs-system. More suitable for comparing OSs are suites like unixbench and lmbench, which include many programs that exercise the OS calls/services. Please find attached the results I have obtained when running the unixbench-4.1.0 for the kernels 2.4.26 and 2.6.14. The 2.4.26 has better indexes for 5 of the 6 tests from unixbench that I have run. On Tuesday 14 March 2006 08:50, David Jander wrote: > On Friday 10 March 2006 16:33, Wolfgang Denk wrote: > > > I believe most of those observations and measurements are not valid > > > anymore. Kernel 2.6 for 8xx has come a long way since this article was > > > written. It might have been true back then, but it surely isn't > > > anymore. > > > > So did you actually run any benchmarks? Specilations on what might be > > or should be are not really helpful. > > Of course I did. Otherwise I wouldn't say this. > > Here's some benchmark data from nbench (sorry didn't try lmbench yet): > > The same ELDK (version 3.1.1) for both kernels, running on exactly the same > board (MPC852T 100MHz, with 32Mbyte SDRAM and 32Mbyte Flash running from > NFS root). I removed some FPU benchmarks, as they are pretty meaningless > for this board and take an ethernity otherwise. Best regards, -- Carlos Mitidieri SYSGO AG - Office Ulm --Boundary-00=_p9qFEshaJsSESxM Content-Type: text/plain; charset="iso-8859-1"; name="unixbench-2.4.26" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="unixbench-2.4.26" BYTE UNIX Benchmarks (Version 4.1.0) System -- Linux 2.4.26 Tue Mar 7 08:56:57 CET 2006 ppc unknown Start Benchmark Run: Thu Jan 1 01:13:27 UTC 1970 1:13am up 1:13, load average: 0.07, 0.73, 0.80 lrwxrwxrwx 1 root root 9 Jan 1 00:00 /bin/sh -> /bin/bash System Call Overhead 243637.1 lps (10.0 secs, 10 samples) Pipe Throughput 246562.1 lps (10.0 secs, 10 samples) Pipe-based Context Switching 114521.0 lps (10.0 secs, 10 samples) Process Creation 2199.0 lps (30.0 secs, 3 samples) Execl Throughput 362.8 lps (29.7 secs, 3 samples) Shell Scripts (1 concurrent) 680.0 lpm (60.0 secs, 3 samples) Shell Scripts (8 concurrent) 92.0 lpm (60.0 secs, 3 samples) Shell Scripts (16 concurrent) 46.0 lpm (60.0 secs, 3 samples) INDEX VALUES TEST BASELINE RESULT INDEX Execl Throughput 43.0 362.8 84.4 Pipe Throughput 12440.0 246562.1 198.2 Pipe-based Context Switching 4000.0 114521.0 286.3 Process Creation 126.0 2199.0 174.5 Shell Scripts (8 concurrent) 6.0 92.0 153.3 System Call Overhead 15000.0 243637.1 162.4 ========= FINAL SCORE 165.8 --Boundary-00=_p9qFEshaJsSESxM Content-Type: text/plain; charset="iso-8859-1"; name="unixbench-2.6.14" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="unixbench-2.6.14" BYTE UNIX Benchmarks (Version 4.1.0) System -- Linux 2.6.14 Tue Mar 7 12:50:28 CET 2006 ppc unknown Start Benchmark Run: Thu Jan 1 00:06:12 UTC 1970 12:06am up 6 min, load average: 0.00, 0.00, 0.00 lrwxrwxrwx 1 root root 9 Jan 1 00:00 /bin/sh -> /bin/bash System Call Overhead 339524.9 lps (10.0 secs, 10 samples) Pipe Throughput 236842.6 lps (10.0 secs, 10 samples) Pipe-based Context Switching 106498.2 lps (10.0 secs, 10 samples) Process Creation 1873.8 lps (30.0 secs, 3 samples) Execl Throughput 372.6 lps (29.8 secs, 3 samples) Shell Scripts (1 concurrent) 650.4 lpm (60.0 secs, 3 samples) Shell Scripts (8 concurrent) 88.0 lpm (60.0 secs, 3 samples) Shell Scripts (16 concurrent) 44.0 lpm (60.0 secs, 3 samples) INDEX VALUES TEST BASELINE RESULT INDEX Execl Throughput 43.0 372.6 86.7 Pipe Throughput 12440.0 236842.6 190.4 Pipe-based Context Switching 4000.0 106498.2 266.2 Process Creation 126.0 1873.8 148.7 Shell Scripts (8 concurrent) 6.0 88.0 146.7 System Call Overhead 15000.0 339524.9 226.3 ========= FINAL SCORE 167.0 --Boundary-00=_p9qFEshaJsSESxM--