From owner-freebsd-stable Sun Jan 9 9:48: 9 2000 Delivered-To: freebsd-stable@freebsd.org Received: from gndrsh.dnsmgr.net (GndRsh.dnsmgr.net [198.145.92.4]) by hub.freebsd.org (Postfix) with ESMTP id 5B86714BED for ; Sun, 9 Jan 2000 09:48:05 -0800 (PST) (envelope-from freebsd@gndrsh.dnsmgr.net) Received: (from freebsd@localhost) by gndrsh.dnsmgr.net (8.9.3/8.9.3) id JAA19344; Sun, 9 Jan 2000 09:47:50 -0800 (PST) (envelope-from freebsd) From: "Rodney W. Grimes" Message-Id: <200001091747.JAA19344@gndrsh.dnsmgr.net> Subject: Re: CPU voltage (was Re: load spike strangeness) In-Reply-To: <3878A3D7.39664AEC@freeloader.freeserve.co.uk> from Ian J Hart at "Jan 9, 2000 03:05:59 pm" To: ianjhart@freeloader.freeserve.co.uk (Ian J Hart) Date: Sun, 9 Jan 2000 09:47:50 -0800 (PST) Cc: rjoseph@nwlink.com (R Joseph Wright), stable@FreeBSD.ORG (stable@freebsd.org) X-Mailer: ELM [version 2.4ME+ PL54 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > R Joseph Wright wrote: > > > Awhile back I was having trouble getting through kernel compiles, the > > machine would reboot during the compile. Someone on some newsgroup said > > "are you overclocking?". I checked my settings. No, I wasn't > > overclocking. But my voltage setting looked wrong. It was set at 2.2 > > volts or whatever and I thought it was supposed to be at 2.4. So I > > changed it to 2.4. My problem was solved. I was flying through the > > compiles. Then I was looking through my motherboard manual and realized > > that it was supposed to be set at 2.2 after all. I prefer it the way it > > is though. Do you think there is anything wrong with that? > > > > -- > > Best Regards, Joseph > > > > You will do foolish things, > > but do them with enthusiasm. Colette. > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > > with "unsubscribe freebsd-stable" in the body of the message > > You don't say what CPU you have. Is it an AMD K6II 400+ by any chance? > > On all the motherbords I've seen lately the CPU PNP for these chips is > broken. They auto detect as 2.4v but have 2.2v stamped on them. > Presumably AMD changed the spec after the BIOS code was written. The > important question is why would they change the core voltage? Temperature? When the K6-2 at 400MHz was first put into production it used the same exact process as the 350Mhz parts and ran at 2.4V. AMD found that the chips where not reliable (so did we, with a close to 30% failure rate) and tweaked the process, this tweaking of the process required a voltage change to 2.2V in the core. I feel sorry for anyone who has a K6-2 400 MHz 2.4V part, they run very hot and often have problems running continuous make worlds for more than a day or so even with a monster heat sink and lots of air flow. > > If anyone is is _really_ interested I can supply voltage and temperature > readings, but not until monday. Maybe I should try 2.3v too. Please post > if you want the data. As another datapoint we recently found, many cooling fans do not seat squarly against the chip due to poor mounting clip design. They often end up with a <50% contact area on the top of the chip, and it is off to the side, away from the die area :-(. The easy way to check for this is to apply light pressure to each of the 4 sides of the heatsink and see if it moves ever so slightly, if it does take it off and look very closely at the pattern in the heat sink compound (if you don't have any of that on there get some unless it has a graphite pad.) I pulled one off a customer had installed the other day and the compound was very thin on one side, very think on the other side, and slightly browned in the thich area. Obvios indication of poor heat transfer. If your fan has a graphite pad on it you can still use the ``wobble'' test to check for this. The problem can often be fixed by simply rotating the fan 180 degrees so that the clip goes on the other way (The poorly designed mounting clips put the pressure point in the center of the clip which is not the center of the chip due to the offset in the socket 7 design. And those clips with the offset in them put on backwards compound the problem by putting the pressure even further from the center of the chip :-()). -- Rod Grimes - KD7CAX @ CN85sl - (RWG25) rgrimes@gndrsh.dnsmgr.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message