From owner-freebsd-questions Mon Nov 1 4: 1:28 1999 Delivered-To: freebsd-questions@freebsd.org Received: from front7.grolier.fr (front7.grolier.fr [194.158.96.57]) by hub.freebsd.org (Postfix) with ESMTP id 05CD514F91 for ; Mon, 1 Nov 1999 04:01:25 -0800 (PST) (envelope-from vons@iname.com) Received: from CYRIL (ppp-116-148.villette.club-internet.fr [194.158.116.148]) by front7.grolier.fr (8.9.3/No_Relay+No_Spam_MGC990224) with ESMTP id NAA04806; Mon, 1 Nov 1999 13:00:45 +0100 (MET) Message-Id: <4.2.1.10.19991101121031.00aa76b8@mail.vons.local> X-Sender: vons@mail.vons.local (Unverified) X-Mailer: QUALCOMM Windows Eudora Pro Version 4.2.1.10 (Beta) Date: Mon, 01 Nov 1999 13:00:26 +0100 To: andrew@sign.chg.ru ("Andrew L. Neporada") From: Gert-Jan Vons Subject: Re: 3.3-STABLE : Performance problems?? Cc: freebsd-questions@FreeBSD.ORG In-Reply-To: <7uvubl$t4u$1@FreeBSD.csie.NCTU.edu.tw> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG At 05:40 1999-10-25 +0800, Andrew L. Neporada wrote: >Hi All! >For a long time I participate in one distributed computational project ( >see http://www.mersenne.org for details). So I'am running a special >program at low priority (nice=20), that utilizes all CPU power. >After upgrading from 3.1-RELEASE to 3.3-STABLE I've noticed that average >performance (averaging period is big enough to eliminate impact from >other tasks ) of this programm is only 77% from measured >performance on 3.1R and same hardware. Same here for the rc5des project (http://www.distributed.net). With FBSD 3.2, I got ~237KKeys/sec. With FBSD3.3, I'm only seeing ~167KKeys/sec, which is only ~70% of the performance I used to see under FBSD3.2 on the same hardware. >I also notice some strange thing: >top shows about 30% 'interrupt' at CPU states line. >Maybe someone could help me to find out what happens? >Any help would be greatly appreciated. I'm not seeing that at all, the interrupt %-age is close to 0% like it used to be. For me, the strange thing is that I see this 30% loss on a Cyrix P166+ rev1.7, but that there is no performance difference whatsoever on an AMD K6-II/300. I haven't been able to find/get an explanation for this :-/ Gert-Jan To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message