clocker
07-06-2004, 03:15 PM
"Are my temps good?"... a question frequently seen here in Hardware world.
Lately I've been thinking about the best way to respond to that and there seem to be two obstacles in the way.
First- what is "good"?
If your machine boots into Windows and runs without errors, then isn't that "good"...or at least "good enough"?
How good is good and at what point does the quest for a lower temp become just silly?
Second- Where are these temps you are reporting coming from, and how much faith should we put in them?
How accurate are the software-derived temps derived from the motherboard and BIOS?
Changing BIOS can radically alter the temp that it reports...which is right?
There does not seem to be a good and universal way to tell.
Recently, I saw a post in which the proposition was put forth that absolute temp ( a number impossible to verify anyway) was basically irrelevant to evaluating a cooling solution.
Of more interest is the difference between case ambient and CPU temp.
Of course, one must assume that both numbers would be equally skewed, but it does seem that this eliminates many variables.
So...if your case temp is say, 25c, and your CPU temp is 40C, then the difference ( or delta figure) is 15c.
This would be considered "good".
If lowering the case temp results in a proportionately lower CPU temp ( in other words, the delta value stays the same or ideally, lowers), then you could assume that the cooling configuration is working properly and your temps are good given the conditions you are operating under.
Discussing our temps like this would seem to eliminate some of the inequities and inaccuracies of just using absolute temp numbers.
After all, if Sprocket was in the desert and the ambient temp was 120 a CPU temp of 135 would actually be pretty good given the circumstances.
If she was in Alaska and the ambient was 50 then 135 CPU temp would totally suck.
So the absolute ( as much as we can trust it) temp we call "good" is variable, while the delta value is not.
What do you think?
Lately I've been thinking about the best way to respond to that and there seem to be two obstacles in the way.
First- what is "good"?
If your machine boots into Windows and runs without errors, then isn't that "good"...or at least "good enough"?
How good is good and at what point does the quest for a lower temp become just silly?
Second- Where are these temps you are reporting coming from, and how much faith should we put in them?
How accurate are the software-derived temps derived from the motherboard and BIOS?
Changing BIOS can radically alter the temp that it reports...which is right?
There does not seem to be a good and universal way to tell.
Recently, I saw a post in which the proposition was put forth that absolute temp ( a number impossible to verify anyway) was basically irrelevant to evaluating a cooling solution.
Of more interest is the difference between case ambient and CPU temp.
Of course, one must assume that both numbers would be equally skewed, but it does seem that this eliminates many variables.
So...if your case temp is say, 25c, and your CPU temp is 40C, then the difference ( or delta figure) is 15c.
This would be considered "good".
If lowering the case temp results in a proportionately lower CPU temp ( in other words, the delta value stays the same or ideally, lowers), then you could assume that the cooling configuration is working properly and your temps are good given the conditions you are operating under.
Discussing our temps like this would seem to eliminate some of the inequities and inaccuracies of just using absolute temp numbers.
After all, if Sprocket was in the desert and the ambient temp was 120 a CPU temp of 135 would actually be pretty good given the circumstances.
If she was in Alaska and the ambient was 50 then 135 CPU temp would totally suck.
So the absolute ( as much as we can trust it) temp we call "good" is variable, while the delta value is not.
What do you think?