Speedfan 4.33

[brassbin]

Do i still need to offset +15C on each core? also do i have to run Prime95 for 24 hours? i oc'ed my E2200's 200Mhz FSB to 266Mhz, when running prime95 both core temp would shot up to 65C at times, would that be considered unstable?

 

[VirtualLarry]

Use CoreTemp 0.96, it is the most accurate for temps. With CoreTemp 0.96.1, or SpeedFan 4.33, you may need to add 15C to core temps.
You should run Prime95 (small FFT) for 24hr for stability testing. Also perhaps the Blend test as well.
65C isn't a problem for load temps. Temp alone doesn't mean unstable. That's what the Prime95 testing is for.

 

[Mondoman]

To clarify, CoreTemp 0.96.1 is the latest version and should work fine with your e2200.

 

[Ika]

It does not, however, work with Vista x64.

 

[brassbin]

I got both CoreTemp 0.96.1 and SpeedFan 4.33 reporting the same temperature, i'll tke it it is reading correctly then. So i've got my E200 FSB to 275Mhz, only 5 seconds into Prime95 Torture test, Core #1 reported hardware failure and the that test thread stopped, the other test thread on Core #2 is still running, and both core temp are around 55C. What does that mean?

 

[brassbin]

ok, both tests stopped, reporting hardware failure, does that mean i need to scale down FSB, and start thinking about raising vcore?

 

[VirtualLarry]

Time to raise vcore then.

 

[brassbin]

Is there a chart detailing correlations between FSB frequency and VCORE value? i assume it's CPU & motherboard specific, but still even if it's only for populars like Gigabyte's P35's, does such a chart exist?

 

[VirtualLarry]

Originally posted by: brassbin
Is there a chart detailing correlations between FSB frequency and VCORE value? i assume it's CPU & motherboard specific, but still even if it's only for populars like Gigabyte's P35's, does such a chart exist?

It's different for each and every CPU. (It's called a schmoo plot in the technical literature.)

That's one of the parts of overclocking, you need to test and graph your own CPU's Freq/Vcore settings.