On Fri, 2003-07-11 at 00:55, Robert Crawford wrote: > On Friday 11 July 2003 12:35 am, Vox wrote: > > On September 1993 plus 3599 days James Sparenberg wrote: > > > Vox, > > > Last ditch if you get worried. open the side and put a small desk fan > > > right on it. From the hardware standpoint. make sure cables (like > > > ribbon cables) are clear of the fan. Also you might consider case fans > > > in addition to the volcano fan. (I love volcano myself.) > > > > I'm not really worried about the temp...having read the docs from > > AMD, I won't worry till it hits 65C, which it has never done :) And > > yes, I like the volcano...has worked well for me...and I do have a > > case fan at the front and a small fan on the mboard itself...I've > > been thinking about getting a second case fan for the back, but am > > too damn lazy to go find one :) > > > > Vox > > It's been my experience with numerous AMD cpus that anything over 45c. is > likely to start causing random misc. problems. I always start feeling > apprehensive when my temps go above 40c. Why AMD states up to 90c. is beyond > me. The one time I got above 50c. the system became unusable- random lockups, > shut-down problems, various hanging with programs, etc. Better cooling fixed > everything instantly, so it was obviously caused by high temps. > > wrc1944 >
This is very true. It's also a fact that higher operating temperatures cause AMD's to operate increasingly hotter as they get older and also shorten the life of the CPU. Once the CPU accumulates damage it cannot be un-damaged. Therefore I made sure that this new XP2100 I got the other day had the best cooling possible, cause I want the processor in mint condition on up to the day it goes on to greener pastures. I also glance with great suspicion at the AMD failure temp. I think that it should be made clearer to users that this is a core temp measurement, which is very different from a mobo sensor temp measurement. There can be 10 to 20 deg C difference between the two, which means you can be frying your CPU and not even realize it. I've got an XP2100 and (it being new) it probably has a core temp diode sensor. But I know for a fact that my system is not taking advantage of that, and that lm_sensors is reporting from the mobo sensor. Therefore to me it's reflexive it to calculate the error of margin in. The small investment in HSF research and hardware is well worth the peace of mind. --LX -- 같같같같같같같같같같같같같같같같같같같같같같같� Linux Mandrake 9.1 Kernel 2.4.21-0.13mdk *Catch Star Trek Enterprise, Wednesdays on UPN* 같같같같같같같같같같같같같같같같같같같같같같같같
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
