When I built my current home computer (what a geeky thing to be able to say), I installed a small applet that sits in the system tray, that little area on the lower right hand part of the screen near the clock. The applet does one thing and one thing alone. It monitors the temperature of my computer’s CPU.
I thought it might be a good idea because sometimes the room gets warm or cool and I wanted to make sure it didn’t suffer. Truth is, My AMD Athlon 1600+ is capable of running at much higher temperatures than what I subject it to. I also thought another cool readout on the screen would be… well… cool.
I was worried about heat, even in the design stages. I have so many fans in the case that it sounds like a Beechcraft 1900 taxiing out for departure.
Give me a sec… I’m getting to the point.
What I found was the biggest contributing factor to higher CPU temperatures was not environmental but actually how much thinking the computer was doing! That was weird., and it took a while to put 2+2 together.
If, for instance, I edit video (which is very math intensive and can take a long time) the CPU’s temperature starts creeping up. On a long session it can be 15-20 degrees warmer than what I normally see. If I’m surfing the net or typing email or working on this blog, it idles relatively cooly.
I’m not sure if this is because in high stress applications there are more electrons trying to move faster, making it a friction thing? The clock that runs the chip keeps a constant beat, so it’s not heating up because its little silicon heart is beating faster.
But, it is a puzzlement.
Here’s what brought this to mind. Tonight, The National Weather Service’s computer, which run the forecast models, had to be shut down because of heat. Since these machines run the same basic programs every night, I don’t think they’re experiencing the same kind of anomolies I see at home. It’s probably mechanical and will need a plumber or air conditioning expert rather than a computer expert. Still, it’s interesting to see that heat is the enemy of computers everywhere.
The link below connects to their statement on the computer failure.
NOUS42 KWNO 262251
SPECIAL NCEP DISCUSSION
CENTRAL OPERATIONS/NOAA/NWS/NCEP WASHINGTON DC
2249 UTC TUE AUG 26 2003
262249Z…NCEP PRODUCTION STOPPED…
THE SNOW SUPERCOMPUTER HAD TO BE SHUT DOWN TO PREVENT
OVERHEATING. AS A RESULT NCEP MODEL PRODUCTION
HAS STOPPED. THE 18Z GFS IS A TOTAL LOSS AND WILL
NOT BE AVAILABLE. BACKUP RUC AND SFC RUC PRODUCTION
HAS BEEN SWITCHED TO BACKUP BEGINNING WITH THE 22Z
ISSUES. THE 21Z RUCS WILL NOT BE AVBL.
THE 00Z ETA AT BEST WILL BE PRODUCED VIA THE BACKUP METHOD.
IT WILL BE AT LEAST A FEW HOURS TILL A SUPERCOMPUTER IS
RETURNED TO SERVICE FOR PRODUCTION. TRAVELERS FCST WILL
NOT BE AVBL. OTHER SIGNIFICANT IMPACTS ARE POSSIBLE.
YOUR PATIENCE AND UNDERSTANDING IS REQUESTED.
GOES WEST WILL BE IN RSO FROM 26/2200Z..27/0700Z
DUE TO FLASH FLOODING/SEVERE WEATHER CONCERNS ACROSS
SOUTHWESTERN US THIS AFTERNOON AND INTO THE OVERNIGHT