From owner-freebsd-smp Sun Sep 19 13: 2: 7 1999 Delivered-To: freebsd-smp@freebsd.org Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (Postfix) with ESMTP id E32BE14D64; Sun, 19 Sep 1999 13:01:59 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.3/8.9.2) with ESMTP id WAA17981; Sun, 19 Sep 1999 22:01:52 +0200 (CEST) (envelope-from phk@critter.freebsd.dk) To: smp@freebsd.org, current@freebsd.org Subject: Testers please! From: Poul-Henning Kamp Date: Sun, 19 Sep 1999 22:01:52 +0200 Message-ID: <17979.937771312@critter.freebsd.dk> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org If you have a PIIX4 based SMP system and run current, could you please try out this patch: http://phk.freebsd.dk/piix/ I'm very interested in hearing if there are any measurable difference apart from clock granularity being 3 times better. -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sun Sep 19 13:49:13 1999 Delivered-To: freebsd-smp@freebsd.org Received: from pop3-3.enteract.com (pop3-3.enteract.com [207.229.143.32]) by hub.freebsd.org (Postfix) with SMTP id 9313415763 for ; Sun, 19 Sep 1999 13:49:09 -0700 (PDT) (envelope-from dscheidt@enteract.com) Received: (qmail 13137 invoked from network); 19 Sep 1999 20:49:09 -0000 Received: from shell-3.enteract.com (dscheidt@207.229.143.42) by pop3-3.enteract.com with SMTP; 19 Sep 1999 20:49:09 -0000 Date: Sun, 19 Sep 1999 15:49:08 -0500 (CDT) From: David Scheidt To: Poul-Henning Kamp Cc: smp@freebsd.org, current@freebsd.org Subject: Re: Testers please! In-Reply-To: <17979.937771312@critter.freebsd.dk> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sun, 19 Sep 1999, Poul-Henning Kamp wrote: > > If you have a PIIX4 based SMP system and run current, could you > please try out this patch: > > http://phk.freebsd.dk/piix/ > > I'm very interested in hearing if there are any measurable difference > apart from clock granularity being 3 times better. What sort of tests would you like done before and after? David Scheidt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sun Sep 19 14: 0:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (Postfix) with ESMTP id DB05D14BCD; Sun, 19 Sep 1999 14:00:34 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.3/8.9.2) with ESMTP id WAA18309; Sun, 19 Sep 1999 22:59:55 +0200 (CEST) (envelope-from phk@critter.freebsd.dk) To: David Scheidt Cc: smp@freebsd.org, current@freebsd.org Subject: Re: Testers please! In-reply-to: Your message of "Sun, 19 Sep 1999 15:49:08 CDT." Date: Sun, 19 Sep 1999 22:59:55 +0200 Message-ID: <18307.937774795@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In message , Dav id Scheidt writes: >On Sun, 19 Sep 1999, Poul-Henning Kamp wrote: > >> >> If you have a PIIX4 based SMP system and run current, could you >> please try out this patch: >> >> http://phk.freebsd.dk/piix/ >> >> I'm very interested in hearing if there are any measurable difference >> apart from clock granularity being 3 times better. > >What sort of tests would you like done before and after? Any test you can think of really. I don't expect any problems, but I am very interested in the performance difference if any, since that would give me a good indication if it is a worthwhile job to spend more time on obfuscat^H^H^H^H^H^H^Hptimizing the timecounter code. As far as I can tell, access to the PIIX timecounter is about 2.5 microseconds faster than to the i8254. In relative terms it is about a factor 3 faster. You can flip forth and back between the old and new behaviour with: New: sysctl -w kern.timecounter.hardware=PIIX (if it refuses you probably don't have the PIIX4 hardware) Old: sysctl -w kern.timecounter.hardware=i8254 -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sun Sep 19 18:24:55 1999 Delivered-To: freebsd-smp@freebsd.org Received: from overcee.netplex.com.au (overcee.netplex.com.au [202.12.86.7]) by hub.freebsd.org (Postfix) with ESMTP id 78C8514BB8; Sun, 19 Sep 1999 18:24:48 -0700 (PDT) (envelope-from peter@netplex.com.au) Received: from netplex.com.au (localhost [127.0.0.1]) by overcee.netplex.com.au (Postfix) with ESMTP id C1C1B1CC5; Mon, 20 Sep 1999 09:24:47 +0800 (WST) (envelope-from peter@netplex.com.au) X-Mailer: exmh version 2.0.2 2/24/98 To: Poul-Henning Kamp Cc: smp@freebsd.org, current@freebsd.org Subject: Re: Testers please! In-reply-to: Your message of "Sun, 19 Sep 1999 22:01:52 +0200." <17979.937771312@critter.freebsd.dk> Date: Mon, 20 Sep 1999 09:24:47 +0800 From: Peter Wemm Message-Id: <19990920012447.C1C1B1CC5@overcee.netplex.com.au> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Poul-Henning Kamp wrote: > > If you have a PIIX4 based SMP system and run current, could you > please try out this patch: > > http://phk.freebsd.dk/piix/ > > I'm very interested in hearing if there are any measurable difference > apart from clock granularity being 3 times better. There is a problem with it as it tries to claim the same device as claimed by pcisupport.c and intpm.c.. pcisupport.c is where some folks have been hanging Tor Egge's RTC SMI trap patch from.. Cheers, -Peter -- Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Sep 20 0:56:18 1999 Delivered-To: freebsd-smp@freebsd.org Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (Postfix) with ESMTP id 27DD215764; Mon, 20 Sep 1999 00:56:11 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.3/8.9.2) with ESMTP id JAA21172; Mon, 20 Sep 1999 09:53:32 +0200 (CEST) (envelope-from phk@critter.freebsd.dk) To: Peter Wemm Cc: smp@freebsd.org, current@freebsd.org Subject: Re: Testers please! In-reply-to: Your message of "Mon, 20 Sep 1999 09:24:47 +0800." <19990920012447.C1C1B1CC5@overcee.netplex.com.au> Date: Mon, 20 Sep 1999 09:53:32 +0200 Message-ID: <21170.937814012@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Ok, noted. I changed to to fail the probe but still use the hardware. Poul-Henning In message <19990920012447.C1C1B1CC5@overcee.netplex.com.au>, Peter Wemm writes : >Poul-Henning Kamp wrote: >> >> If you have a PIIX4 based SMP system and run current, could you >> please try out this patch: >> >> http://phk.freebsd.dk/piix/ >> >> I'm very interested in hearing if there are any measurable difference >> apart from clock granularity being 3 times better. > >There is a problem with it as it tries to claim the same device as claimed >by pcisupport.c and intpm.c.. pcisupport.c is where some folks have been >hanging Tor Egge's RTC SMI trap patch from.. > >Cheers, >-Peter >-- >Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au > > -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Sep 20 5:41:43 1999 Delivered-To: freebsd-smp@freebsd.org Received: from overcee.netplex.com.au (overcee.netplex.com.au [202.12.86.7]) by hub.freebsd.org (Postfix) with ESMTP id CDD0714CEF; Mon, 20 Sep 1999 05:41:29 -0700 (PDT) (envelope-from peter@netplex.com.au) Received: from netplex.com.au (localhost [127.0.0.1]) by overcee.netplex.com.au (Postfix) with ESMTP id 509B71CC5; Mon, 20 Sep 1999 20:41:19 +0800 (WST) (envelope-from peter@netplex.com.au) X-Mailer: exmh version 2.0.2 2/24/98 To: Poul-Henning Kamp Cc: smp@freebsd.org, current@freebsd.org Subject: Re: Testers please! In-reply-to: Your message of "Mon, 20 Sep 1999 09:53:32 +0200." <21170.937814012@critter.freebsd.dk> Date: Mon, 20 Sep 1999 20:41:19 +0800 From: Peter Wemm Message-Id: <19990920124119.509B71CC5@overcee.netplex.com.au> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Poul-Henning Kamp wrote: > > Ok, noted. I changed to to fail the probe but still use the hardware. If intpm is probed first (the smbus driver), your probe won't even get called. I think we need an early quirks or hooks handler in the pci probes to handle stuff like this. For example, we have hooks fixing up a handful of wierd bios misconfigurations, collecting these together via a quirks table or whatever would also give a convenient place for you to hook this sort of thing into, and without it being dependent on link or probe order. > Poul-Henning > > In message <19990920012447.C1C1B1CC5@overcee.netplex.com.au>, Peter Wemm writ es > : > >Poul-Henning Kamp wrote: > >> > >> If you have a PIIX4 based SMP system and run current, could you > >> please try out this patch: > >> > >> http://phk.freebsd.dk/piix/ > >> > >> I'm very interested in hearing if there are any measurable difference > >> apart from clock granularity being 3 times better. > > > >There is a problem with it as it tries to claim the same device as claimed > >by pcisupport.c and intpm.c.. pcisupport.c is where some folks have been > >hanging Tor Egge's RTC SMI trap patch from.. Cheers, -Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Sep 20 5:59:49 1999 Delivered-To: freebsd-smp@freebsd.org Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (Postfix) with ESMTP id 1434F14F1D; Mon, 20 Sep 1999 05:59:42 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.3/8.9.2) with ESMTP id OAA22738; Mon, 20 Sep 1999 14:56:54 +0200 (CEST) (envelope-from phk@critter.freebsd.dk) To: Peter Wemm Cc: smp@freebsd.org, current@freebsd.org Subject: Re: Testers please! In-reply-to: Your message of "Mon, 20 Sep 1999 20:41:19 +0800." <19990920124119.509B71CC5@overcee.netplex.com.au> Date: Mon, 20 Sep 1999 14:56:54 +0200 Message-ID: <22736.937832214@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In message <19990920124119.509B71CC5@overcee.netplex.com.au>, Peter Wemm writes : >Poul-Henning Kamp wrote: >> >> Ok, noted. I changed to to fail the probe but still use the hardware. > >If intpm is probed first (the smbus driver), your probe won't even get >called. I think we need an early quirks or hooks handler in the pci probes >to handle stuff like this. For example, we have hooks fixing up a handful >of wierd bios misconfigurations, collecting these together via a quirks >table or whatever would also give a convenient place for you to hook this >sort of thing into, and without it being dependent on link or probe order. sigh... -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Sep 20 18:44:32 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mail.rdc1.sfba.home.com (ha1.rdc1.sfba.home.com [24.0.0.66]) by hub.freebsd.org (Postfix) with ESMTP id 30BDA14C89 for ; Mon, 20 Sep 1999 18:44:31 -0700 (PDT) (envelope-from adsharma@c62443-a.frmt1.sfba.home.com) Received: from c62443-a.frmt1.sfba.home.com ([24.0.69.165]) by mail.rdc1.sfba.home.com (InterMail v4.01.01.00 201-229-111) with ESMTP id <19990921014430.GLFH29487.mail.rdc1.sfba.home.com@c62443-a.frmt1.sfba.home.com> for ; Mon, 20 Sep 1999 18:44:30 -0700 Received: (from adsharma@localhost) by c62443-a.frmt1.sfba.home.com (8.9.3/8.9.3) id SAA16575 for freebsd-smp@freebsd.org; Mon, 20 Sep 1999 18:44:30 -0700 Date: Mon, 20 Sep 1999 18:44:30 -0700 From: Arun Sharma To: freebsd-smp@freebsd.org Subject: Context switch code structure Message-ID: <19990920184430.A16551@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.6i Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org After the recent change to move the "chooseproc" logic from assembly to C, I wondered if the code should be restructured in a different way. Right now, it approximately looks like: proc = chooseproc Now, if one day FreeBSD achieves fine grained locking and there are multiple CPUs in the kernel mode, which means certain conditions can change over time, underneath the current process. Specifically, if the current process is the most eligible process after context switching out, then we switch to the same process wasting a bunch of CPU cycles. Are there any guarantees that this can't happen ? If not, would it make sense to restructure the algorithm to do: proc = chooseproc switch(current, proc) -Arun To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 12:25:28 1999 Delivered-To: freebsd-smp@freebsd.org Received: from hromeo.algonet.se (hromeo.algonet.se [194.213.74.10]) by hub.freebsd.org (Postfix) with SMTP id 6A66C14F39 for ; Tue, 21 Sep 1999 12:25:23 -0700 (PDT) (envelope-from mal@algonet.se) Received: (qmail 14850 invoked from network); 21 Sep 1999 21:25:21 +0200 Received: from kent.algonet.se (194.213.74.90) by hromeo.algonet.se with SMTP; 21 Sep 1999 21:25:21 +0200 Received: from kairos.algonet.se ([194.213.74.18]) by algonet.se (BLUETAIL Mail Robustifier1.0.4) with ESMTP ; Tue, 21 Sep 1999 19:25:21 GMT Received: (mal@localhost) by kairos.algonet.se (8.8.8+Sun/8.6.12) id VAA29450; Tue, 21 Sep 1999 21:25:20 +0200 (MET DST) Date: Tue, 21 Sep 1999 21:25:20 +0200 (MET DST) Message-Id: <199909211925.VAA29450@kairos.algonet.se> X-Authentication-Warning: kairos.algonet.se: mal set sender to mal@kairos.algonet.se using -f From: Mats Lofkvist To: adams@digitalspark.net Cc: somsky@annwn.phys.washington.edu, freebsd-smp@FreeBSD.ORG In-reply-to: (message from Adam Strohl on Thu, 16 Sep 1999 20:58:38 +0000 (GMT)) Subject: Re: Examples of FreeBSD SMP success? References: Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Adam Strohl wrote: Running two copies of seti@home results in twice the number of blocks being cleared in the same amount of time as 1. How large caches do you have on that machine?!? On my kind-of-dated dual pentium pro with 256k caches a seti@home block takes 24.6 hours on one cpu, but when running two in parallell they need 43.9 hours each. So the speedup going from one to two processors is a mere 12%. Easier tasks as rc5des or parallell makes achive close to factor two speedups. (I'm running current, cvsupped and rebuilt august 14.) _ Mats Lofkvist mal@algonet.se To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 12:41:14 1999 Delivered-To: freebsd-smp@freebsd.org Received: from c004.sfo.cp.net (c004-h006.c004.sfo.cp.net [209.228.14.77]) by hub.freebsd.org (Postfix) with SMTP id 4DC4215E96 for ; Tue, 21 Sep 1999 12:41:11 -0700 (PDT) (envelope-from jason@intercom.com) Received: (cpmta 15803 invoked from network); 21 Sep 1999 12:41:10 -0700 Received: from shagalicious.com (HELO intercom.com) (206.98.165.250) by smtp.intercom.com with SMTP; 21 Sep 1999 12:41:10 -0700 X-Sent: 21 Sep 1999 19:41:10 GMT Message-ID: <37E7DF8C.1EA6B133@intercom.com> Date: Tue, 21 Sep 1999 15:42:04 -0400 From: "Jason J. Horton" Organization: Intercom Online Inc. X-Mailer: Mozilla 4.61 [en] (X11; I; FreeBSD 3.3-STABLE i386) X-Accept-Language: en MIME-Version: 1.0 To: freebsd-smp@FreeBSD.ORG Subject: Abit BP6 References: <199909211925.VAA29450@kairos.algonet.se> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Has anyone run across problems using an Abit BP6 on -STABLE? Was nice of Abit to give FreeBSD a mention in the motherboard manual. -- -Jason J. Horton Moving Target Intercom Online Inc. 212.376.7440 ext 21 | http://www.intercom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 13:54:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mercury.gfit.net (ns.gfit.net [209.41.124.90]) by hub.freebsd.org (Postfix) with ESMTP id 36FA315A91 for ; Tue, 21 Sep 1999 13:51:12 -0700 (PDT) (envelope-from tom@embt.com) Received: from paranor.embt.net (timembt.iinc.com [206.67.169.229]) by mercury.gfit.net (8.8.8/8.8.8) with SMTP id PAA01862 for ; Tue, 21 Sep 1999 15:55:14 -0500 (CDT) (envelope-from tom@embt.com) Message-Id: <3.0.3.32.19990921165107.00939384@mail.embt.com> X-Sender: tembt@mail.embt.com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Tue, 21 Sep 1999 16:51:07 -0400 To: freebsd-smp@freebsd.org From: Tom Embt Subject: Re: Abit BP6 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 03:42 PM 9/21/99 -0400, you wrote: >Has anyone run across problems using an Abit BP6 on -STABLE? > >Was nice of Abit to give FreeBSD a mention in the motherboard manual. > I ran -stable on my BP6 for a short while, now it's running -current. No problems to speak of with either branch, although I'm awaiting a driver for the HighPoint controller (status on this, anyone? anyone?). Also, isn't it a nice change to see FreeBSD listed _before_ Linux in the BP6 manual! [whoops, I cc'd to -stable instead of -smp, now sending to -smp as I should've, forgive the crosspost] Tom Embt ICQ UIN: 11245398 tom@embt.com d:-)> ------------------------------------------------------------------ "You're one of those condescending Unix computer users!" "Here's a nickel, kid. Get yourself a better computer" - Dilbert. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 14: 7:19 1999 Delivered-To: freebsd-smp@freebsd.org Received: from freebsd.dk (freebsd.dk [212.242.42.178]) by hub.freebsd.org (Postfix) with ESMTP id 4208114BEC for ; Tue, 21 Sep 1999 14:07:13 -0700 (PDT) (envelope-from sos@freebsd.dk) Received: (from sos@localhost) by freebsd.dk (8.9.3/8.9.1) id XAA21369; Tue, 21 Sep 1999 23:06:52 +0200 (CEST) (envelope-from sos) From: Soren Schmidt Message-Id: <199909212106.XAA21369@freebsd.dk> Subject: Re: Abit BP6 In-Reply-To: <3.0.3.32.19990921165107.00939384@mail.embt.com> from Tom Embt at "Sep 21, 1999 04:51:07 pm" To: tom@embt.com (Tom Embt) Date: Tue, 21 Sep 1999 23:06:52 +0200 (CEST) Cc: freebsd-smp@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL54 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org It seems Tom Embt wrote: > > At 03:42 PM 9/21/99 -0400, you wrote: > >Has anyone run across problems using an Abit BP6 on -STABLE? > > > >Was nice of Abit to give FreeBSD a mention in the motherboard manual. > > > > I ran -stable on my BP6 for a short while, now it's running -current. No > problems to speak of with either branch, although I'm awaiting a driver for > the HighPoint controller (status on this, anyone? anyone?). Working on it, I've just been promised a controller based on the HPT366 chip.... -Søren To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 14:51:11 1999 Delivered-To: freebsd-smp@freebsd.org Received: from pike.sover.net (pike.sover.net [209.198.87.34]) by hub.freebsd.org (Postfix) with ESMTP id 23540152A5 for ; Tue, 21 Sep 1999 14:51:05 -0700 (PDT) (envelope-from adams@digitalspark.net) Received: from nightfall.digitalspark.net (arc0a326.bf.sover.net [209.198.83.234]) by pike.sover.net (8.9.3/8.9.3) with ESMTP id RAA10941; Tue, 21 Sep 1999 17:50:55 -0400 (EDT) Comments: SoVerNet Verification (on pike.sover.net) nightfall.digitalspark.net from arc0a326.bf.sover.net [209.198.83.234] 209.198.83.234 Tue, 21 Sep 1999 17:50:55 -0400 (EDT) Date: Tue, 21 Sep 1999 17:54:48 +0000 (GMT) From: Adam Strohl To: Mats Lofkvist Cc: somsky@annwn.phys.washington.edu, freebsd-smp@FreeBSD.ORG Subject: Re: Examples of FreeBSD SMP success? In-Reply-To: <199909211925.VAA29450@kairos.algonet.se> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I have dual Celeron 300As clocked to 450Mhz (100Mhz x4.5), 128k of cache per CPU running at 450Mhz. It takes around 5 hours on one CPU to clear a block, and litterally the same time to do two at the same time (I tested it, I can't remeber the numbers exactly, but they were like a second off). - ----( Adam Strohl )------------------------------------------------ - - UNIX Operations/Systems http://www.digitalspark.net - - adams (at) digitalspark.net xxx.xxx.xxxx xxxxx - - ----------------------------------------( DigitalSpark.NET )------- - On Tue, 21 Sep 1999, Mats Lofkvist wrote: > Adam Strohl wrote: > > Running two copies of seti@home results in twice the number of blocks > being cleared in the same amount of time as 1. > > How large caches do you have on that machine?!? > On my kind-of-dated dual pentium pro with 256k caches a seti@home block > takes 24.6 hours on one cpu, but when running two in parallell they need > 43.9 hours each. So the speedup going from one to two processors is a > mere 12%. Easier tasks as rc5des or parallell makes achive close to factor > two speedups. > > (I'm running current, cvsupped and rebuilt august 14.) > > _ > Mats Lofkvist > mal@algonet.se > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 15:15:25 1999 Delivered-To: freebsd-smp@freebsd.org Received: from post.mail.nl.demon.net (post-11.mail.nl.demon.net [194.159.73.21]) by hub.freebsd.org (Postfix) with ESMTP id 41C5015687 for ; Tue, 21 Sep 1999 15:15:18 -0700 (PDT) (envelope-from marc@oldserver.demon.nl) Received: from [212.238.105.241] (helo=mistress) by post.mail.nl.demon.net with esmtp (Exim 2.12 #1) id 11TYDX-000LKg-00; Tue, 21 Sep 1999 22:16:55 +0000 Date: Wed, 22 Sep 1999 00:14:28 +0200 (CEST) From: Marc Schneiders To: Adam Strohl Cc: Mats Lofkvist , somsky@annwn.phys.washington.edu, freebsd-smp@freebsd.org Subject: Re: Examples of FreeBSD SMP success? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 21 Sep 1999, Adam Strohl wrote: > I have dual Celeron 300As clocked to 450Mhz (100Mhz x4.5), 128k of cache > per CPU running at 450Mhz. > > It takes around 5 hours on one CPU to clear a block, and litterally the > same time to do two at the same time (I tested it, I can't remeber the > numbers exactly, but they were like a second off). > > - ----( Adam Strohl )------------------------------------------------ - > - UNIX Operations/Systems http://www.digitalspark.net - > - adams (at) digitalspark.net xxx.xxx.xxxx xxxxx - > - ----------------------------------------( DigitalSpark.NET )------- - > > On Tue, 21 Sep 1999, Mats Lofkvist wrote: > > > Adam Strohl wrote: > > > > Running two copies of seti@home results in twice the number of blocks > > being cleared in the same amount of time as 1. > > > > How large caches do you have on that machine?!? > > On my kind-of-dated dual pentium pro with 256k caches a seti@home block > > takes 24.6 hours on one cpu, but when running two in parallell they need > > 43.9 hours each. So the speedup going from one to two processors is a > > mere 12%. Easier tasks as rc5des or parallell makes achive close to factor > > two speedups. > > > > (I'm running current, cvsupped and rebuilt august 14.) > > > > _ > > Mats Lofkvist > > mal@algonet.se > > > > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-smp" in the body of the message > How on earth do you do it? Softupdates?? My UniProcessor Celeron366 @ 458 takes 15 hours (4.0 current of 13 September). I do use it for mail and the like, but no CPU intensive stuff. I have also run FreeBSD (both 3.1 and 4.0) on a Dual PPro 200 @ 233. One Seti file was only slightly quicker than letting the box do two at a time. Maybe 5 % or so. Of course two CPU's need twice the amount of RAM, I would say. But I hardly expect anyone running an SMP box with less than 64 MB. So that cannot be the issue, probably. Marc marc@oldserver.demon.nl To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 15:52:18 1999 Delivered-To: freebsd-smp@freebsd.org Received: from pike.sover.net (pike.sover.net [209.198.87.34]) by hub.freebsd.org (Postfix) with ESMTP id EA52C1545D for ; Tue, 21 Sep 1999 15:52:13 -0700 (PDT) (envelope-from adams@digitalspark.net) Received: from nightfall.digitalspark.net (arc0a326.bf.sover.net [209.198.83.234]) by pike.sover.net (8.9.3/8.9.3) with ESMTP id SAA28775; Tue, 21 Sep 1999 18:52:03 -0400 (EDT) Comments: SoVerNet Verification (on pike.sover.net) nightfall.digitalspark.net from arc0a326.bf.sover.net [209.198.83.234] 209.198.83.234 Tue, 21 Sep 1999 18:52:03 -0400 (EDT) Date: Tue, 21 Sep 1999 18:55:55 +0000 (GMT) From: Adam Strohl To: Marc Schneiders Cc: Mats Lofkvist , somsky@annwn.phys.washington.edu, freebsd-smp@freebsd.org Subject: Re: Examples of FreeBSD SMP success? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Wed, 22 Sep 1999, Marc Schneiders wrote: > How on earth do you do it? Softupdates?? My UniProcessor Celeron366 @ 458 > takes 15 hours (4.0 current of 13 September). I do use it for mail and the > like, but no CPU intensive stuff. > I have also run FreeBSD (both 3.1 and 4.0) on a Dual PPro 200 @ 233. One > Seti file was only slightly quicker than letting the box do two at a time. > Maybe 5 % or so. Of course two CPU's need twice the amount of RAM, I > would say. But I hardly expect anyone running an SMP box with less than 64 > MB. So that cannot be the issue, probably. Hmmm, Asus P2B-D motherboard, with 128MB of RAM, IDE disks. seti is VERY light on disk. Also, I meant to say 15 hours, not 5. However, it took 15 hours if 1 block was processing, or 15 hours if there were two. That was my point. - ----( Adam Strohl )------------------------------------------------ - - UNIX Operations/Systems http://www.digitalspark.net - - adams (at) digitalspark.net xxx.xxx.xxxx xxxxx - - ----------------------------------------( DigitalSpark.NET )------- - To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 16: 1:17 1999 Delivered-To: freebsd-smp@freebsd.org Received: from horst.bfd.com (horst.bfd.com [12.9.219.10]) by hub.freebsd.org (Postfix) with ESMTP id D248F14EF0 for ; Tue, 21 Sep 1999 16:01:15 -0700 (PDT) (envelope-from ejs@bfd.com) Received: from HARLIE.bfd.com (bastion.bfd.com [12.9.219.14]) by horst.bfd.com (8.9.3/8.9.2) with ESMTP id QAA07766; Tue, 21 Sep 1999 16:01:14 -0700 (PDT) (envelope-from ejs@bfd.com) Date: Tue, 21 Sep 1999 16:01:14 -0700 (PDT) From: "Eric J. Schwertfeger" To: "Jason J. Horton" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Abit BP6 In-Reply-To: <37E7DF8C.1EA6B133@intercom.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 21 Sep 1999, Jason J. Horton wrote: > Has anyone run across problems using an Abit BP6 on -STABLE? Not -STABLE, but both 3.2 and 3.3 -RELEASE. (O.K., so I've only been running 3.3 for 4 days). To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 17:25:14 1999 Delivered-To: freebsd-smp@freebsd.org Received: from lor.watermarkgroup.com (lor.watermarkgroup.com [207.202.73.33]) by hub.freebsd.org (Postfix) with ESMTP id DD8A914BD8; Tue, 21 Sep 1999 17:25:09 -0700 (PDT) (envelope-from luoqi@watermarkgroup.com) Received: (from luoqi@localhost) by lor.watermarkgroup.com (8.8.8/8.8.8) id UAA25595; Tue, 21 Sep 1999 20:25:08 -0400 (EDT) (envelope-from luoqi) Date: Tue, 21 Sep 1999 20:25:08 -0400 (EDT) From: Luoqi Chen Message-Id: <199909220025.UAA25595@lor.watermarkgroup.com> To: current@FreeBSD.ORG, phk@FreeBSD.ORG, smp@FreeBSD.ORG Subject: Re: Testers please! Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > If you have a PIIX4 based SMP system and run current, could you > please try out this patch: > > http://phk.freebsd.dk/piix/ > > I'm very interested in hearing if there are any measurable difference > apart from clock granularity being 3 times better. > > -- > Poul-Henning Kamp FreeBSD coreteam member > phk@FreeBSD.ORG "Real hackers run -current on their laptop." > FreeBSD -- It will take a long time before progress goes too far! > This reminds me about the usage of TSC counter on SMP. First even though we don't use TSC for time keeping on SMP, the TSC frequency from calibration is still valid (at least for BSP), and we can show it in the cpu identification message. Second, the listed reason for not using TSC on SMP is the inability to synchronize TSCs on all cpus. My question is, is it really necessary? TSC is initialized to 0 at hardware reset, which should happen to all CPUs at the same time (invalid assumption?), in another words, all TSCs should be automatically synchronized. -lq To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 20:13:58 1999 Delivered-To: freebsd-smp@freebsd.org Received: from janus.syracuse.net (janus.syracuse.net [205.232.47.15]) by hub.freebsd.org (Postfix) with ESMTP id 58ECF14EF0; Tue, 21 Sep 1999 20:13:39 -0700 (PDT) (envelope-from green@FreeBSD.org) Received: from localhost (green@localhost) by janus.syracuse.net (8.9.3/8.8.7) with ESMTP id XAA14018; Tue, 21 Sep 1999 23:13:42 -0400 (EDT) X-Authentication-Warning: janus.syracuse.net: green owned process doing -bs Date: Tue, 21 Sep 1999 23:13:36 -0400 (EDT) From: "Brian F. Feldman" X-Sender: green@janus.syracuse.net To: Luoqi Chen Cc: current@FreeBSD.org, phk@FreeBSD.org, smp@FreeBSD.org Subject: Re: Testers please! In-Reply-To: <199909220025.UAA25595@lor.watermarkgroup.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 21 Sep 1999, Luoqi Chen wrote: > This reminds me about the usage of TSC counter on SMP. First even though > we don't use TSC for time keeping on SMP, the TSC frequency from calibration > is still valid (at least for BSP), and we can show it in the cpu identification > message. Second, the listed reason for not using TSC on SMP is the inability > to synchronize TSCs on all cpus. My question is, is it really necessary? > TSC is initialized to 0 at hardware reset, which should happen to all CPUs > at the same time (invalid assumption?), in another words, all TSCs should be > automatically synchronized. ISTRT something I did long ago was kill the TSC reset FreeBSD did, so there should be no reason that they won't be all at the same place. I am willing to bet half my farm that the biggest problem was that we did that, and barring that we could have always done SMP using the TSC. One question comes to mind: is there a way that the TSCs could become desynchronized somehow? Even though all CPUs run at the same frequency, isn't there a strong possibility for slight frequency deviation since we use crystal oscillation instead of a more accurate atomic breakdown for regulation, or am I just smoking crack? > > -lq > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-current" in the body of the message > -- Brian Fundakowski Feldman \ FreeBSD: The Power to Serve! / green@FreeBSD.org `------------------------------' To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 20:22:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from shell.webmaster.com (mail.webmaster.com [209.133.28.73]) by hub.freebsd.org (Postfix) with ESMTP id 33CF414F4D; Tue, 21 Sep 1999 20:22:33 -0700 (PDT) (envelope-from davids@webmaster.com) Received: from whenever ([209.133.29.2]) by shell.webmaster.com (Post.Office MTA v3.5.3 release 223 ID# 0-12345L500S10000V35) with SMTP id com; Tue, 21 Sep 1999 20:22:32 -0700 From: "David Schwartz" To: "Brian F. Feldman" , "Luoqi Chen" Cc: , Subject: RE: Testers please! Date: Tue, 21 Sep 1999 20:22:32 -0700 Message-ID: <000001bf04a9$b50ebf90$021d85d1@youwant.to> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook 8.5, Build 4.71.2377.0 In-Reply-To: Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > One question comes to mind: is there a way that the TSCs could become > desynchronized somehow? Even though all CPUs run at the same frequency, > isn't there a strong possibility for slight frequency deviation since > we use crystal oscillation instead of a more accurate atomic breakdown > for regulation, or am I just smoking crack? All CPUs should be clocked off of the same frequency multiplier off of the same crystal, so it _should_ be impossible for the TSCs to drift apart due to that. I guess it might be theoretically possible that something might cause the TSCs to drift. Perhaps something crazy with APM and/or SMM. It should be tested. It should also be tested that they begin at the same place. Perhaps processors might take different amount of times to 'come out of' reset. I wouldn't assume anything. DS To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 21: 0:59 1999 Delivered-To: freebsd-smp@freebsd.org Received: from dingo.cdrom.com (dingo.cdrom.com [204.216.28.145]) by hub.freebsd.org (Postfix) with ESMTP id 3F8BF15B4C; Tue, 21 Sep 1999 21:00:31 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost.cdrom.com [127.0.0.1]) by dingo.cdrom.com (8.9.3/8.8.8) with ESMTP id UAA10049; Tue, 21 Sep 1999 20:53:33 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Message-Id: <199909220353.UAA10049@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: "Brian F. Feldman" Cc: Luoqi Chen , current@FreeBSD.org, phk@FreeBSD.org, smp@FreeBSD.org Subject: Re: Testers please! In-reply-to: Your message of "Tue, 21 Sep 1999 23:13:36 EDT." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 21 Sep 1999 20:53:33 -0700 From: Mike Smith Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > One question comes to mind: is there a way that the TSCs could become > desynchronized somehow? Even though all CPUs run at the same frequency, > isn't there a strong possibility for slight frequency deviation since > we use crystal oscillation instead of a more accurate atomic breakdown > for regulation, or am I just smoking crack? They should all be using the same clock source, so that's not an issue. -- \\ The mind's the standard \\ Mike Smith \\ of the man. \\ msmith@freebsd.org \\ -- Joseph Merrick \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Sep 21 22:56:25 1999 Delivered-To: freebsd-smp@freebsd.org Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (Postfix) with ESMTP id ECB9C14D83; Tue, 21 Sep 1999 22:56:21 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.3/8.9.2) with ESMTP id HAA32779; Wed, 22 Sep 1999 07:55:37 +0200 (CEST) (envelope-from phk@critter.freebsd.dk) To: Luoqi Chen Cc: current@FreeBSD.ORG, smp@FreeBSD.ORG Subject: Re: Testers please! In-reply-to: Your message of "Tue, 21 Sep 1999 20:25:08 EDT." <199909220025.UAA25595@lor.watermarkgroup.com> Date: Wed, 22 Sep 1999 07:55:36 +0200 Message-ID: <32777.937979736@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In message <199909220025.UAA25595@lor.watermarkgroup.com>, Luoqi Chen writes: >Second, the listed reason for not using TSC on SMP is the inability >to synchronize TSCs on all cpus. My question is, is it really necessary? Strictly speaking no, it isn't necessary, but unless they are in sync the timekeeping code gets very complex. See Dave "ntp" Mills code to hold them in sync. >TSC is initialized to 0 at hardware reset, which should happen to all CPUs >at the same time (invalid assumption?), in another words, all TSCs should be >automatically synchronized. They are not. The PLL is local to each cpu and every single clock-stop/start event has then inching away from each other because the on-chip VCO is very temperature dependent. Furthermore Linux people have found sufficiently many cases where the counters are not in sync after the BIOS is done. -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Sep 22 6:14:34 1999 Delivered-To: freebsd-smp@freebsd.org Received: from host.phc.igs.net (host.phc.igs.net [216.58.103.2]) by hub.freebsd.org (Postfix) with ESMTP id 42B5F14D26; Wed, 22 Sep 1999 06:14:30 -0700 (PDT) (envelope-from eagle@phc.igs.net) Received: from ttyA0f.phc.igs.net (ttyA0f.phc.igs.net [216.58.103.45]) by host.phc.igs.net (8.8.5/8.8.5) with ESMTP id JAA17437; Wed, 22 Sep 1999 09:25:36 -0400 (EDT) Date: Wed, 22 Sep 1999 09:15:27 -0400 (EDT) From: Rob Garrett To: David Schwartz Cc: "Brian F. Feldman" , Luoqi Chen , phk@FreeBSD.ORG, smp@FreeBSD.ORG Subject: RE: Testers please! In-Reply-To: <000001bf04a9$b50ebf90$021d85d1@youwant.to> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 21 Sep 1999, David Schwartz wrote: > > > One question comes to mind: is there a way that the TSCs could become > > desynchronized somehow? Even though all CPUs run at the same frequency, > > isn't there a strong possibility for slight frequency deviation since > > we use crystal oscillation instead of a more accurate atomic breakdown > > for regulation, or am I just smoking crack? > > All CPUs should be clocked off of the same frequency multiplier off of the > same crystal, so it _should_ be impossible for the TSCs to drift apart due > to that. I guess it might be theoretically possible that something might > cause the TSCs to drift. Perhaps something crazy with APM and/or SMM. It > should be tested. > > It should also be tested that they begin at the same place. Perhaps > processors might take different amount of times to 'come out of' reset. > > I wouldn't assume anything. > > DS > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-smp" in the body of the message > > since the spec requires the ap kick the bp's alive, they won't come out of reset at the same time. the bp sits and waits, for the ap, to tell it to come alive. Rob To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Sep 22 7:39:33 1999 Delivered-To: freebsd-smp@freebsd.org Received: from lor.watermarkgroup.com (lor.watermarkgroup.com [207.202.73.33]) by hub.freebsd.org (Postfix) with ESMTP id 5552A15240; Wed, 22 Sep 1999 07:39:28 -0700 (PDT) (envelope-from luoqi@watermarkgroup.com) Received: (from luoqi@localhost) by lor.watermarkgroup.com (8.8.8/8.8.8) id KAA05687; Wed, 22 Sep 1999 10:39:27 -0400 (EDT) (envelope-from luoqi) Date: Wed, 22 Sep 1999 10:39:27 -0400 (EDT) From: Luoqi Chen Message-Id: <199909221439.KAA05687@lor.watermarkgroup.com> To: luoqi@watermarkgroup.com, phk@critter.freebsd.dk Subject: Re: Testers please! Cc: current@FreeBSD.ORG, smp@FreeBSD.ORG Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > >TSC is initialized to 0 at hardware reset, which should happen to all CPUs > >at the same time (invalid assumption?), in another words, all TSCs should be > >automatically synchronized. > > They are not. The PLL is local to each cpu and every single > clock-stop/start event has then inching away from each other because > the on-chip VCO is very temperature dependent. Furthermore Linux The internal clock is phase-locked to an external clock source, which should be the same bus clock and identical for all cpus. The multipliers for the internal clocks should also be identical (in almost all cases unless you design your own mother board). The highest multiplier as of today is 6x (600MHz cpu on a 100MHz bus), it might increase over time, but I expect it to be still around 10. If the PLL's accuracy is 1%, the difference between two cpus could at most be 10% (or 3% = 1% * sqrt(10), assuming a Gaussian distribution of the drift) of the internal clock cycle, that is 10% (3%) of the time you might get two different readings (differ by 1) if you read TSCs of two cpus simultaneously, that's still much more accurate than the i8254 timer. > people have found sufficiently many cases where the counters are > not in sync after the BIOS is done. > I would really like to know how they managed to read the TSCs simultaneously, or they resorted to some kind of statistical means (which I tried without much success, maybe I will try later with some kernel hooks). The multiplier is set at hard reset, so BIOS couldn't change that (some mother boards may be soft configurable, but the change has to take effect after next reset). Bus clock is external, BIOS couldn't change that either. So the internal clock is not affected by BIOS. Now it would be really hard or me to believe that TSC is not a simple counter that increments at each internal clock edge. I am very skeptical about the Linux people's findings. > -- > Poul-Henning Kamp FreeBSD coreteam member > phk@FreeBSD.ORG "Real hackers run -current on their laptop." > FreeBSD -- It will take a long time before progress goes too far! > -lq To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Sep 22 7:52: 5 1999 Delivered-To: freebsd-smp@freebsd.org Received: from lor.watermarkgroup.com (lor.watermarkgroup.com [207.202.73.33]) by hub.freebsd.org (Postfix) with ESMTP id DD8B314EEF; Wed, 22 Sep 1999 07:52:02 -0700 (PDT) (envelope-from luoqi@watermarkgroup.com) Received: (from luoqi@localhost) by lor.watermarkgroup.com (8.8.8/8.8.8) id KAA05897; Wed, 22 Sep 1999 10:51:05 -0400 (EDT) (envelope-from luoqi) Date: Wed, 22 Sep 1999 10:51:05 -0400 (EDT) From: Luoqi Chen Message-Id: <199909221451.KAA05897@lor.watermarkgroup.com> To: davids@webmaster.com, eagle@phc.igs.net Subject: RE: Testers please! Cc: green@FreeBSD.ORG, luoqi@watermarkgroup.com, phk@FreeBSD.ORG, smp@FreeBSD.ORG Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > since the spec requires the ap kick the bp's alive, they won't come out of reset > at the same time. the bp sits and waits, for the ap, to tell it to come alive. > > Rob > First, I think you ap->BSP, bp->AP. Second, the spec says, the APs sit and wait for SIPI in a halted state after losing out BIPI. The spec also says, TSC is incremented even after the processor is halted by the HLT instruction, which I believe is the same "halted" state. -lq To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Sep 22 8: 8: 3 1999 Delivered-To: freebsd-smp@freebsd.org Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (Postfix) with ESMTP id 3F53915442; Wed, 22 Sep 1999 08:07:55 -0700 (PDT) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.3/8.9.2) with ESMTP id RAA34338; Wed, 22 Sep 1999 17:07:10 +0200 (CEST) (envelope-from phk@critter.freebsd.dk) To: Luoqi Chen Cc: current@FreeBSD.ORG, smp@FreeBSD.ORG Subject: Re: Testers please! In-reply-to: Your message of "Wed, 22 Sep 1999 10:39:27 EDT." <199909221439.KAA05687@lor.watermarkgroup.com> Date: Wed, 22 Sep 1999 17:07:10 +0200 Message-ID: <34336.938012830@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In message <199909221439.KAA05687@lor.watermarkgroup.com>, Luoqi Chen writes: >> people have found sufficiently many cases where the counters are >> not in sync after the BIOS is done. >> >I would really like to know how they managed to read the TSCs simultaneously, >or they resorted to some kind of statistical means (which I tried without >much success, maybe I will try later with some kernel hooks). The differences were pretty significant in offset, I havn't heard any numbers from them on skew, and I don't know of anybody who have tried to measure it. The trouble with N counters for N>1 is that you need to add code to synchronize AND syntonize them, you need code to make sure time is monotonic no matter what and yada yada yada. By the time you're done the i8254 doesn't look so bad after all... The ACPI counter is a good sized step in the right direction, they should have made it PCI clock driven, to get better resolution but I suspect the fact that they didn't is a sign that the PCI clock will be power munged at some point. For "general purpose" time a clock which is about the resolution of the IO busses is sufficient, the ACPI is a little low on frequency, but hey, it's miles better than the i8254, and the APM/ACPI will not fuck with it's frequency. For high resolution timing of "in-cpu" events, the TSC is still the way to go, and nothing (except SMP complexity) prevents it from being used for that, and it is probably pointless to convert the data to nanoseconds until the difference has been found in the end anyway. -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Sep 22 12:20:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from host.phc.igs.net (host.phc.igs.net [216.58.103.2]) by hub.freebsd.org (Postfix) with ESMTP id 79415159C6 for ; Wed, 22 Sep 1999 12:20:12 -0700 (PDT) (envelope-from eagle@phc.igs.net) Received: from [216.58.103.28] ([216.58.103.28]) by host.phc.igs.net (8.8.5/8.8.5) with ESMTP id PAA26076 for ; Wed, 22 Sep 1999 15:31:23 -0400 (EDT) Date: Wed, 22 Sep 1999 15:21:20 -0400 (EDT) From: Rob Garrett To: smp@freebsd.org Subject: patches to disable smi y2k interupts Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org phk, suggested that i ask here, where i could find such monstrosities, it appears that the smi interupt, is causing problems in two places. any ideas where to find the patch? robg To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Sep 24 0: 4:16 1999 Delivered-To: freebsd-smp@freebsd.org Received: from dingo.cdrom.com (castles558.castles.com [208.214.165.122]) by hub.freebsd.org (Postfix) with ESMTP id A4D8A14F9C for ; Fri, 24 Sep 1999 00:04:14 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost [127.0.0.1]) by dingo.cdrom.com (8.9.3/8.8.8) with ESMTP id XAA01804; Thu, 23 Sep 1999 23:56:24 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Message-Id: <199909240656.XAA01804@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Rob Garrett Cc: smp@freebsd.org Subject: Re: patches to disable smi y2k interupts In-reply-to: Your message of "Wed, 22 Sep 1999 15:21:20 EDT." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 23 Sep 1999 23:56:24 -0700 From: Mike Smith Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > phk, suggested that i ask here, where i could find such monstrosities, it > appears that the smi interupt, is causing problems in two places. any ideas > where to find the patch? Update your BIOS, don't futz with the intended operation of your system. (There's a known bug in at least the 1008 family of BIOS revisions from ASUS that cause these symptoms. Poul appears to have missed the notes that upgrading to 1010 fixes these.) -- \\ The mind's the standard \\ Mike Smith \\ of the man. \\ msmith@freebsd.org \\ -- Joseph Merrick \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Sep 24 1:13:45 1999 Delivered-To: freebsd-smp@freebsd.org Received: from host.phc.igs.net (host.phc.igs.net [216.58.103.2]) by hub.freebsd.org (Postfix) with ESMTP id 50D7814D71 for ; Fri, 24 Sep 1999 01:13:42 -0700 (PDT) (envelope-from eagle@phc.igs.net) Received: from ttyA0b.phc.igs.net (ttyA0b.phc.igs.net [216.58.103.41]) by host.phc.igs.net (8.8.5/8.8.5) with ESMTP id EAA09049; Fri, 24 Sep 1999 04:23:23 -0400 (EDT) Date: Fri, 24 Sep 1999 04:13:22 -0400 (EDT) From: Rob Garrett To: Mike Smith Cc: smp@freebsd.org Subject: Re: patches to disable smi y2k interupts In-Reply-To: <199909240656.XAA01804@dingo.cdrom.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org interesting, I'm using the 1010 version of the bios, on the asus, the problems are sio overflow and negative calcru errors. rob On Thu, 23 Sep 1999, Mike Smith wrote: > > > > phk, suggested that i ask here, where i could find such monstrosities, it > > appears that the smi interupt, is causing problems in two places. any ideas > > where to find the patch? > > Update your BIOS, don't futz with the intended operation of your system. > > (There's a known bug in at least the 1008 family of BIOS revisions > from ASUS that cause these symptoms. Poul appears to have missed the > notes that upgrading to 1010 fixes these.) > > -- > \\ The mind's the standard \\ Mike Smith > \\ of the man. \\ msmith@freebsd.org > \\ -- Joseph Merrick \\ msmith@cdrom.com > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Sep 24 21: 3: 1 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mailer.tydfam.machida.tokyo.jp (ns.tydfam.machida.tokyo.jp [210.161.209.130]) by hub.freebsd.org (Postfix) with ESMTP id E7C0A14D4B for ; Fri, 24 Sep 1999 21:02:56 -0700 (PDT) (envelope-from ken@tydfam.machida.tokyo.jp) Received: from localhost (ns1.tydfam.machida.tokyo.jp [210.161.209.138]) by mailer.tydfam.machida.tokyo.jp (8.9.3/3.7W10/03/98) with ESMTP id NAA00342 for ; Sat, 25 Sep 1999 13:02:55 +0900 (JST) To: freebsd-smp@FreeBSD.ORG Subject: Q) current SMP on ASUS P65UP/C-P6ND X-Mailer: Mew version 1.93 on XEmacs 20.4 (Emerald) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <19990925130254Y.ken@ns1.tydfam.machida.tokyo.jp> Date: Sat, 25 Sep 1999 13:02:54 +0900 From: Takeshi Yamada X-Dispatcher: imput version 980905(IM100) Lines: 14 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Recent cvsup on current source caused my PPro x2 box (Asus P65UP/C-P6ND) locks in a few minute after startup, and does not work. It was OK with about Aug. 15th update. Recent ML does not show any such problems, and makes me suspect if it is only Socket-8 box issue. Are there any fixes on going? or do I need to give up obsolete Socket-8 board? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Sep 24 21: 7:54 1999 Delivered-To: freebsd-smp@freebsd.org Received: from southpass.baynetworks.com (ns2.BayNetworks.COM [134.177.3.16]) by hub.freebsd.org (Postfix) with ESMTP id 8D7D014BC6 for ; Fri, 24 Sep 1999 21:07:53 -0700 (PDT) (envelope-from thomma@BayNetworks.COM) Received: from mailhost.BayNetworks.COM (h016b.s86b1.BayNetworks.COM [134.177.1.107]) by southpass.baynetworks.com (8.9.1/8.9.1) with ESMTP id VAA25755; Fri, 24 Sep 1999 21:02:47 -0700 (PDT) Received: from fedex.engwest.baynetworks.com (fedex.engwest.baynetworks.com [134.177.110.46]) by mailhost.BayNetworks.COM (8.9.1/8.8.8) with SMTP id VAA21180; Fri, 24 Sep 1999 21:06:37 -0700 (PDT) Received: from carrera.engwest (carrera.engwest.baynetworks.com) by fedex.engwest.baynetworks.com (4.1/SMI-4.1) Received: from localhost by carrera.engwest (SMI-8.6/SMI-SVR4) id VAA21294; Fri, 24 Sep 1999 21:04:33 -0700 To: ken@tydfam.machida.tokyo.jp Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Q) current SMP on ASUS P65UP/C-P6ND In-Reply-To: Your message of "Sat, 25 Sep 1999 13:02:54 +0900" <19990925130254Y.ken@ns1.tydfam.machida.tokyo.jp> References: <19990925130254Y.ken@ns1.tydfam.machida.tokyo.jp> X-Mailer: Mew version 1.92 on Emacs 19.28 / Mule 2.3 (SUETSUMUHANA) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <19990924210433P.thomma@baynetworks.com> Date: Fri, 24 Sep 1999 21:04:33 -0700 From: Tamiji Homma X-Dispatcher: imput version 971024 Lines: 15 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Recent cvsup on current source caused my PPro x2 box > (Asus P65UP/C-P6ND) locks in a few minute after startup, > and does not work. > > It was OK with about Aug. 15th update. Recent ML does > not show any such problems, and makes me suspect if it is > only Socket-8 box issue. > > Are there any fixes on going? or do I need to give up > obsolete Socket-8 board? I haven't seen a single problem recently (last update 24 hour old) -current with 2 Seti@home clients/buildworld loop. Tammy To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Sep 25 2:13:33 1999 Delivered-To: freebsd-smp@freebsd.org Received: from dingo.cdrom.com (castles557.castles.com [208.214.165.121]) by hub.freebsd.org (Postfix) with ESMTP id 4860514D36 for ; Sat, 25 Sep 1999 02:13:29 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost [127.0.0.1]) by dingo.cdrom.com (8.9.3/8.8.8) with ESMTP id CAA01432; Sat, 25 Sep 1999 02:05:50 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Message-Id: <199909250905.CAA01432@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Takeshi Yamada Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Q) current SMP on ASUS P65UP/C-P6ND In-reply-to: Your message of "Sat, 25 Sep 1999 13:02:54 +0900." <19990925130254Y.ken@ns1.tydfam.machida.tokyo.jp> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sat, 25 Sep 1999 02:05:50 -0700 From: Mike Smith Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org If you have APM enabled, disable it. > Recent cvsup on current source caused my PPro x2 box > (Asus P65UP/C-P6ND) locks in a few minute after startup, > and does not work. > > It was OK with about Aug. 15th update. Recent ML does > not show any such problems, and makes me suspect if it is > only Socket-8 box issue. > > Are there any fixes on going? or do I need to give up > obsolete Socket-8 board? -- \\ The mind's the standard \\ Mike Smith \\ of the man. \\ msmith@freebsd.org \\ -- Joseph Merrick \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Sep 25 10:12:18 1999 Delivered-To: freebsd-smp@freebsd.org Received: from macs.mxim.com (macs.mxim.com [204.17.143.130]) by hub.freebsd.org (Postfix) with ESMTP id F121915193 for ; Sat, 25 Sep 1999 10:11:58 -0700 (PDT) (envelope-from michaele@mxim.com) Received: from localhost (michaele@localhost) by macs.mxim.com (8.7/8.6.9) with ESMTP id KAA01386 for ; Sat, 25 Sep 1999 10:11:56 -0700 (PDT) Date: Sat, 25 Sep 1999 10:11:56 -0700 (PDT) From: Michael Enkelis X-Sender: michaele@macs To: freebsd-smp@freebsd.org Subject: PNP gone Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Does anyone know when the "pnp" command in userconfig will be restored? I use it to setup my TYAN motherboards builtin sound chip. _ _ _ __ michaele@mxim.com ' ) ) ) / /) / ` / /) Michael Enkelis / / / o _. /_ __. _ // /-- __ /_ _ // o _ (503) 641 - 3737 x2245 / ' (_(_(__/ /_(_(_(<_(/_ (___, /) )_/ <_(<_(/_(_/_)_ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Sep 25 20:22:52 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mailer.tydfam.machida.tokyo.jp (ns.tydfam.machida.tokyo.jp [210.161.209.130]) by hub.freebsd.org (Postfix) with ESMTP id 3797C15381 for ; Sat, 25 Sep 1999 20:22:36 -0700 (PDT) (envelope-from ken@tydfam.machida.tokyo.jp) Received: from localhost (ns1.tydfam.machida.tokyo.jp [210.161.209.138]) by mailer.tydfam.machida.tokyo.jp (8.9.3/3.7W10/03/98) with ESMTP id MAA03704; Sun, 26 Sep 1999 12:22:25 +0900 (JST) To: mike@smith.net.au Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Q) current SMP on ASUS P65UP/C-P6ND Reply-To: ken@tydfam.machida.tokyo.jp In-Reply-To: Your message of "Sat, 25 Sep 1999 02:05:50 -0700" <199909250905.CAA01432@dingo.cdrom.com> References: <199909250905.CAA01432@dingo.cdrom.com> X-Mailer: Mew version 1.93 on XEmacs 20.4 (Emerald) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <19990926122225P.ken@ns1.tydfam.machida.tokyo.jp> Date: Sun, 26 Sep 1999 12:22:25 +0900 From: Takeshi Yamada X-Dispatcher: imput version 980905(IM100) Lines: 64 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Thank you for your quick response. From: Tamiji Homma > I haven't seen a single problem recently (last update 24 hour > old) -current with 2 Seti@home clients/buildworld loop. From: Mike Smith mike> mike> If you have APM enabled, disable it. mike> No, I do not enable APM and functions it does not have like USB. I zapped /usr/src/* and made 'cvs checkout' so that I make sure the source is the ratest one, and rebuild kernel using update LINT minus lines I don't need. It starts up without any problem to somewhere /etc/rc.local, and went to message saying 'mp_lock ......'. The followings are kgdb, but it seems like not usable as it says that dump at kern_shutdow.c. I never asked the machine to shutdown at booting up. Also, kernel with another config file I have been using for sometime caused freeze of I/O, but not core dump a few minutes after booting. No dump was available. Any suggestions for further dig up? -------------------------------------------------------------- (kgdb) where #0 boot (howto=256) at ../../kern/kern_shutdown.c:281 #1 0xc01679e4 in poweroff_wait (junk=0xc02b5c20, howto=-1070900278) at ../../kern/kern_shutdown.c:531 #2 0xc018dd6f in lookup (ndp=0xc612293c) at ../../kern/vfs_lookup.c:528 #3 0xc018d533 in namei (ndp=0xc612293c) at ../../kern/vfs_lookup.c:152 #4 0xc0156fb2 in elf_load_file (p=0xc576f780, file=0xc61229f0 "/usr/libexec/ld-elf.so.1", addr=0xc61229ec, entry=0xc6122eb4) at ../../kern/imgact_elf.c:331 #5 0xc0157701 in exec_elf_imgact (imgp=0xc6122e88) at ../../kern/imgact_elf.c:589 #6 0xc01606ae in execve (p=0xc576f780, uap=0xc6122f80) at ../../kern/kern_exec.c:179 #7 0xc027cd62 in syscall (frame={tf_fs = 47, tf_es = 47, tf_ds = 47, tf_edi = 0, tf_esi = 134885440, tf_ebp = -1077944904, tf_isp = -971886636, tf_ebx = -1077944968, tf_edx = -1077944957, tf_ecx = 134901831, tf_eax = 59, tf_trapno = 12, tf_err = 2, tf_eip = 134559312, tf_cs = 31, tf_eflags = 662, tf_esp = -1077945008, tf_ss = 47}) at ../../i386/i386/trap.c:1056 #8 0xc026a4d1 in Xint0x80_syscall () (kgdb) list 276 EVENTHANDLER_INVOKE(shutdown_post_sync, howto); 277 splhigh(); 278 if ((howto & (RB_HALT|RB_DUMP)) == RB_DUMP && !cold) { 279 savectx(&dumppcb); 280 #ifdef __i386__ 281 dumppcb.pcb_cr3 = rcr3(); 282 #endif 283 dumpsys(); 284 } 285 (kgdb) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message