Reasoning for multiple Prime passes, then fail?

S

supastar1568

Limp Gawd
Joined
Jul 19, 2006
Messages
470
Hey folks,

Recently clocked my e8400 to 3.8 Ghz. Tested and passed: 21 hours Prime Small FFT, 9 hours Large FFT, and 25 hours blend.

Today, randomly I decided to run small FFT again. This time it failed after 26 minutes.

What's the reasoning behind this? Ambient temps are roughly the same.
 
Hmm thats a very odd case, did you try turning off your comp for a few hours then trying again later? I mean maybe its just a fluke? How long ago did you fully test your oc? Were you running any extra programs (was your cpu usage close to 0% before you turned the test on)? Did the max temp change compared to before? How did u measure ambient temperature, I mean did you use a thermometer or does it simply feels the same?
 
Maybe your CPU has suffered some degradation or 'burn-in' as some people like to say. You may need to either increase your CPU voltage or drop your speed down.
 
I have not taken the CPU over 1.325 Volts, that's my limit.

Maybe the ambient temps were a bit off, not sure, but my room does get hot, about 30C ambient (90 degrees F). It could indeed be a fluke, I will try again overnight. Interesting stuff. Should not be a problem hitting 3.8 GHz. I may just try a 423 FSB clock with a 9x multi, instead of 475 and 8x.
 
Not always, when you are talking about computer, ambient temp can be the case temp, which is always hotter than or at best equal to room temp.

Normally ambient temp is room temp.
 
Roomtemp is a consistent measure we can use to see if the CPU temp is within spec.
Different peoples casetemp varies for the same roomtemp so it isnt a usable metric.
 
Just a thought though, I am not saying you are wrong or anything, I am just wondering if this is the case.
Wouldn't that give you more reason to use case temp instead of room temp if you need to knwo if the CPU temp is within spec for the reason you just gave? Say for example, if your cpu is at 50C load, and room temp is 5c, you would think the CPU is NOT within spec, however a closer examination revealed that the case temp is 40c, then you would say wow, that's pretty good temp. This is an extreme example, but do you get my idea? Using the case temp would actually give you a more meaningful measure for temp in your computer.
The exact reason why you want to use case temp would be the reason why you said one shouldn't, it varies for different people under the same room temp, therefore you should use that. Since all temp inside your computer varies depend on the case temp.
 
Same thing happened to me on my E8400. Turned out to be bad RAM that memtest86 was missing. Stable as a rock now.
 
Saggy said:
Just a thought though, I am not saying you are wrong or anything, I am just wondering if this is the case.
Wouldn't that give you more reason to use case temp instead of room temp if you need to knwo if the CPU temp is within spec for the reason you just gave? Say for example, if your cpu is at 50C load, and room temp is 5c, you would think the CPU is NOT within spec, however a closer examination revealed that the case temp is 40c, then you would say wow, that's pretty good temp. This is an extreme example, but do you get my idea? Using the case temp would actually give you a more meaningful measure for temp in your computer.
The exact reason why you want to use case temp would be the reason why you said one shouldn't, it varies for different people under the same room temp, therefore you should use that. Since all temp inside your computer varies depend on the case temp.
Click to expand...

Case airflow also affects case temp so its not as easy to work with.
We do sometimes ask if the case has decent airflow, but there isnt a need to know the actual case temp for our purposes in troubleshooting.
 
You must log in or register to reply here.
Back
Top