Message boards : Number crunching : Temperature really DOES affect results.
Author | Message |
---|---|
Chilean Send message Joined: 16 Oct 05 Posts: 711 Credit: 26,694,507 RAC: 0 |
My little jewel (AMD Athlon 64 x2 6400+) (I run on a limited budget) was spewing out plenty of errors @ 2,000 secs and above. I thought it was Rosetta's fault. Until I found out that the temps were on the neighborhood of 58-65. So I took the time to take out the heatsink (an antec, those big round heatsinks), and got the dust out by blowing on it... and let me tell you, those darn things collect more dust than a vacuum does. But anyways, I placed new thermal paste, and hooked it on the CPU. Then I blew air inside the case at random places to get rid of dust bunnies that like to play and hide within my electronics. (I also got the pleasant opportunity of accidentally eating a few of them while at it). Then I hooked everything back again. Launched BOINC, and it stayed at 50C on 100%. That's an almost a 15C decrease! And I've only gotten a single error so far (a 0.000 sec error... which is Rosetta's fault). :D Moral of story: High temps screw with your CPU. |
Evan Send message Joined: 23 Dec 05 Posts: 268 Credit: 402,585 RAC: 0 |
The Christmas Lectures here in the UK put on by the Royal Institution covered the subject of computers. As part of it they viewed a processor with a thermal imaging camera and it was amazing to see how hot it gets. I had a look around on the net and found this article written in 2005 about this subject. http://www.bigbruin.com/reviews05/nf4thermals/index.php?file=1 |
dcdc Send message Joined: 3 Nov 05 Posts: 1832 Credit: 119,688,048 RAC: 9,222 |
it'll use more power when hot too... |
Message boards :
Number crunching :
Temperature really DOES affect results.
©2024 University of Washington
https://www.bakerlab.org