Yep, I gave kernel-smp-26-2.6.1-5.x86_64.rpm a try and got a kernel mode
panic, something about the file system. I'm using an lsi logic 320-2
megaraid controller, suse is installed on a resier file system on an array
off the lsi logic.
Cheers.
-----Original Message-----
From: Andreas Jaeger [mailto:aj@suse.de]
Sent: Sunday, January 25, 2004 2:04 PM
To: Jason Clark
Subject: Re: [suse-amd64] Quad opteron performance issues in suse 9
"Jason Clark"
Are there any 2.6 kernels available for suse 9, pre-built? Haven't had a whole lot of luck building one.
Yes, there are - in ftp.suse.com/pub/people/kraxel Andreas
Cheers
-----Original Message----- From: Andreas Jaeger [mailto:aj@suse.de] Sent: Sunday, January 25, 2004 9:37 AM To: Jason Clark Cc: suse-amd64@suse.com Subject: Re: [suse-amd64] Quad opteron performance issues in suse 9
"Jason Clark"
writes: I've been playing around with Suse 9 AMD64 for the past few days, trying some canned benchmarks and one that we've constructed. So far, I'm not seeing any scaling from a dual opteron to a quad opteron on suse 9. Most of the tests I've tried are based off of mysql. During the quad cpu tests each cpu was averaging at least 90%, same as the dual cpu tests. Is the 2.4 kernel ok for quad opterons, or is 2.6 required?
The kernel is fine for 4 ways and it scales with our experiments.
The 2.6 kernel is far less locks than 2.4 so depending on your tests, updating might make a difference.
A side issue, for some reason you cannot install the rpm binaries of
mysql
4.0.17 without getting a segmentation fault. So all of the mysql testing
Please tell that to those that did the rpms...
was done on 4.0.15, included with suse 9.
Andreas
Andreas -- Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj SuSE Linux AG, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126