A Prime95 question

Hi folks,

Got my EVGA Superclock cooler installed and started working on overclocking my E8500.

I'm currently running Prime95 64bit version 26.6. It's been running for nearly 10 hours without any apparent problem, but the strange thing (to me, anyway) is that "worker 1" is ahead of "worker 2." What I mean is, "worker 1" was already on Test 10 of the 1M length test when "worker 2" started Test 1.

Is this normal or should I be concerned? Could it be some Windows "housekeeping" task that has caused this unsynchronicity?

During preliminary testing, while searching for the correct Vcore setting, it was always core1 ("worker 2") that would error out.

Anyhow, I may call it good after 12 hours and post some photos and benchmarks.

I don't have the answer, but running Prime on my i5-2550K, 4 cores, one core will run 6-8 tests out of synch, sometime they catch up though. Run a temp monitor with Prime to see if the offending core is getting hotter than the other. It may be just throttling back on frequency and voltage.

Britton30 said:

I don't have the answer, but running Prime on my i5-2550K, 4 cores, one core will run 6-8 tests out of synch, sometime they catch up though. Run a temp monitor with Prime to see if the offending core is getting hotter than the other. It may be just throttling back on frequency and voltage.

Good to know that others have experienced this as well. In that case, I won't let it trouble me too much.

It did run for 12 hours without problems, so I called it good for now.

Got my E8500 up from 3.16GHz to 4.27GHz, I think that's fantastic!

I'm going to run with this until it quits or becomes so outdated that I can't stand it any longer.

Temps were good (if just a tad high) with only 1 degree between the cores.

Yeah, CoreTemp is in Fahrenheit just so I can relate to the number, but I'm becoming more comfortable with Celsius.

Very good OC chap! It would bet best to rum only one temp monitoring software when benchmarking, it can cause them to give errant readings.
another stress test we use a lot is Intel Burn Test, whic can bring a CPU to its knees. If you can pass it at standard your should be good to go. http://www.xtremesystems.org/forums/...esting-program

Britton30 said:

Very good OC chap! It would bet best to rum only one temp monitoring software when benchmarking, it can cause them to give errant readings.
another stress test we use a lot is Intel Burn Test, whic can bring a CPU to its knees. If you can pass it at standard your should be good to go. IntelBurnTest - The new stress-testing program

I hear ya about the multiple temp monitors. I started CoreTemp only to show the CPU speed. :)

Doesn't IBT use LinPack? I started a LinPack test in OCCT and the temps went through the roof immediately.

Fired up my favorite game (Borderlands) with a temp monitor on the secondary monitor so I could see it during gameplay, temps stayed at a reasonable level.

It doesn't show in the HWiNFO capture in my previous post, but I also have a slight overclock on my graphics card, from 900MHz to 950MHz on GPU clock and from 2206MHz to 2400MHz on memory clock.

I'll try IBT and see what happens...hopefully not an earth shattering kaboom.

Lava King said:

Doesn't IBT use LinPack? I started a LinPack test in OCCT and the temps went through the roof immediately.

Thats correct : both IBT and OCCT use Linpack, and yes the temperatures should spike very quickly.

Heres a little tutorial on IBT if it helps:

CPU - Stress Test Using IntelBurnTest

Regards,
Golden