On Tue, 4 Dec 2001 at 18:09, plug bert wrote: > Intel processors have a contingency plan for overheating problems -- > they just lower the Mhz's down until the temp becomes acceptable(some > sort of thermal diode thingy, but only on P4's i think).
Hey, this is neat-o! > Now, before u feel that Intel is cheating u out of the Mhz's , consider > what AMD processors do when they overheat: they just hang up on you? > Your data? Goes to /dev/null B) Processor overheats are a dangerous thing. You get things from plain old hangups to sig 11's to kernel panics to, God forbid, some other hardware component of your system going down permanently together with your heating CPU. And in a country like ours you need to not just have good system ventillation, you need to be in an airconditioned room. > But hey, this is just my personal experience. If u still remain an AMD > fan, then at the very least put the system through its paces: runs > benchmarks and drive that cpu up to 100% utilization. If there are any > AMD success stories out there, please do share your experiences! Make me > an AMD convert B) No matter what processor you pick, I think it's important to do two things. First is configure a hardware monitoring program. lm_sensors is great, and together with rrdtool has very good web-based CGIs for graphs with decent history (one week to view with summary page). Unfortunately you may have some wierd hardware whose manufacturer is a bane to the computing industry (ie: refuses to release the specs, forcing the lm_sensors team to do brute force reverse engineering, an example is ASUS who until now refuses to release specs of sensors like those I have), so YMMV. But for most situations the documentation is good enough to work with. Next is to run something to keep your CPU at 100% usage for at least a few days. On Linux, cpuburn is a program that's meant to do exactly this. Together with temperature monitoring, you can take this opportunity (while you hopefully still have that warranty card intact) to check how your system "scales" with continuous CPU usage. For the positivist, the numbers a properly configured lm_sensors installation will give in a test like this is evidence enough. Better than any speculation anyone else can throw. --> Jijo -- Federico Sevilla III :: [EMAIL PROTECTED] Network Administrator :: The Leather Collection, Inc. GnuPG Key: <http://jijo.leathercollection.ph/jijo.gpg> _ Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph To leave: send "unsubscribe" in the body to [EMAIL PROTECTED] To subscribe to the Linux Newbies' List: send "subscribe" in the body to [EMAIL PROTECTED]
