From owner-freebsd-stable Fri Jun 4 11:22: 9 1999 Delivered-To: freebsd-stable@freebsd.org Received: from pollux.sdata.de (pollux.sdata.de [193.30.133.37]) by hub.freebsd.org (Postfix) with ESMTP id 990A115AD4 for ; Fri, 4 Jun 1999 11:22:03 -0700 (PDT) (envelope-from cs@sdata.de) Received: from sdata.de (vega.sdata.de [193.30.133.36]) by pollux.sdata.de (8.9.3/8.9.2) with ESMTP id UAA44726; Fri, 4 Jun 1999 20:20:45 +0200 (CEST) (envelope-from cs@sdata.de) Message-ID: <37581908.76DF506C@sdata.de> Date: Fri, 04 Jun 1999 20:20:56 +0200 From: Christoph Splittgerber Organization: sdata - C. Splittgerber Datentechnik X-Mailer: Mozilla 4.6 [en] (X11; I; FreeBSD 3.2-STABLE i386) X-Accept-Language: en MIME-Version: 1.0 To: John Reynolds~ Cc: freebsd-stable@freebsd.org Subject: Re: Bug in SMP Kernel in 3.1, was it fixed in 3.2? References: <19990603104448.24898.rocketmail@web118.yahoomail.com> <3756F682.9CA5E2B3@sdata.de> <3757A071.422E1984@wtb.tue.nl> <3757F57B.AE8A68B8@sdata.de> <14167.63562.830344.476388@hip186.ch.intel.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG John Reynolds~ wrote: > Other than that wierd problem with 'top' does SMP work OK on the P2B-DS? I > bought this motherboard about 2 months ago. I only have one PII in it now > but was planning on putting the other in sometime towards the end of the > summer. > > Do you think it's some thing with top, kernel data structures, or do you > believe the 2nd CPU is slowly being "shut off" (i.e. can you do something > really intense like "make -j 2 buildworld" w/ and w/o SMP and measure the > wall-clock time?). > > What BIOS do you have loaded in there. I recently upgraded to BIOS 1009 > (which contains an upgrade of adaptec's scsi bios to v2.11) and it helped > me actually install 3.1-RELEASE on this box (haven't had time to CVSup to > 3.2-STABLE yet). > John, I really don't know about the source of the problem. It's just the %CPU and %WCPU that decrease to 0.00 a short while after booting. The process time is accumulated ok I think. Yes, I'm sure that both CPU's are running - I have 2 "setiathome jobs" (see http://setiathome.ssl.berkeley.edu) niced at -19 running all the time in the background. These jobs are absolutely cpu intensive and take 100% of the CPU if they can. One job finishes in about 10h, 2 jobs also finish in about 10h and in top I see the run state changes between RUN, CPU0, CPU1. I recognized one other strange behaviour: If I have these to seti jobs running disk access becomes very slow - even if I "idprio" them down as far as possible. A simple "ls -ls" may take a couple of seconds in a directory with only 100 entries. If I stop just one of the background jobs, disk access is fast as ever. I don't see this on my other single-cpu FreeBSD machines, they all run a seti-job in the background and besides taking some memory they don't slow down anything when running at nice -19. However, all single-cpu hosts have a AHA2940(UW) controller, so maybe this is specific to the aic7890/91 chipset and not a SMP problem at all. The BIOS here is Rev. 1008 (I think) so maybe I should get an update. However, besides these two minor problems, the machine runs without any problems since 2 month now. Christoph To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message