Hi, I have a number of developers running a mix of Suse 10.2 and Suse 10.0. These developers are all running a Java application called 'maven' that runs unit tests on our code and also automatically creates installers. This system also uses MySQL pretty heavily. I have had a lot of complaints from the 10.2 users that their maven runs are significantly (up to 50%) slower than the 10.0 machines. All machines are of a similiar spec. So in order to investigate, I installed a dual boot of 10.0 and 10.2 (x86) on a 3Ghz box with 1Gb RAM. Both installs used Reiser for the filesystem. I ran the maven app on both and saw 10.0 completed in 29 minutes and 10.2 completed in 61 minutes. I then went a bit deeper and ran bonnie (disk benchtest) on both. This showed the following, Suse 10.0 ---Sequential Output (nosync)--- ---Sequential Input-- --Rnd Seek- -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --04k (03)- Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU maven0 1*1000 40748 99.6105847 22.9 28052 7.7 28561 55.0 61974 9.9 744.9 2.2 Suse 10.2 ---Sequential Output (nosync)--- ---Sequential Input-- --Rnd Seek- -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --04k (03)- Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU maven2 1*1000 30340 97.2 90815 20.1 25670 6.1 19712 74.1 66955 11.2 466.3 1.2 The times for the maven run (unit tests of application source and product builds using maven) are, (not identical to the first run, but withing the same magnitude), Suse 10.0 real 27m5.559s user 13m34.899s sys 0m57.340s Suse 10.2 real 73m19.454s user 13m8.509s sys 0m40.863s The uname -a for suse 10.0 is linux 2.6.13-15-smp #1 SMP Tue Sep 13 14:56:15 UTC 2005 i686 i686 i386 GNU/Linux The uname -a for suse 10.2 is linux 2.6.18.2-34-default #1 SMP Mon Nov 27 11:46:27 UTC 2006 i686 i686 i386 GNU/Linux Can anyone shed any light or have any suggestions? Please find attached the hardware information - siga.txt.bz2 Regards, Ian Collins.