View Full Version : Correlation between folding and temps
Arkaine23
10-17-02, 06:15 PM
I was playing with gromacs lately, and in searching thru my EM3 log I noticed something. My computers are in the garage, and so ambient temps really affect them. A cold front came through two days ago, and my times on P180's were netter by a full minute/frame on those two days, and my load temp went down to 37C from about 43C. And when I fired up a Gromacs WU today, the first half dozen frames were pretty quick, then the heat of the day came on and my CPU temp shot up by 5C. The Gromacs WU's time/frame grew longer by 30 seconds, 45 seconds, and then a minute and 15 seconds!!! Then it hit 50C and crashed the core78. :( Anyway, just seemed odd that an Athlon was affected like this while folding.
john240sx
10-17-02, 06:42 PM
intresting discovery. i wouldn't think that the CPU temp would affect it's performance as long as it's stabil (that is of course until you hit supercooling temps)
i'll have to check this out on my computer and see if it does that same thing.
NASsoccer
10-17-02, 09:53 PM
really odd that is would make that much of a difference before. maybe the rig is stable but the core itself is not. so yes, the rig seems to be going fine but the core itself does not like working with the high heat and as the temp rises, instability occurs and then the crash eventually happens?!?!
FOLD ON
NAS
overdoze
10-17-02, 09:57 PM
There is a new version of EIII 2.0.1 go and d/l it it may help. I'm going to do it now
Arkaine23
10-17-02, 10:01 PM
The main thingy I wanteed to point out is not that core78 crashed when my cpu temps started going up in the heat of the day and slowed down as it became more unstable... although that illustrates the reverse. But that during a cold front, for 2 days on the Tinker core, my times on P180's were 7:30/frame, whereas for weeks before that in my log they were always more like 8:30 or even 8:40 and that at reasonable 43-45C cpu temps.
overdoze
10-17-02, 10:16 PM
that is strange. what I could think off is the code of the program try to turn off the SSE for a moment then turn it back on due to the -forceasm when it detect instability. High temp could make the cpu core unstable especially the sse modules as it is much more complicate than the FPU and ALU. I'm folding with cpu temp at 50C today as well but the difference is I'm watercooling.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.