PassMark reports different CPU speed then BIOS

[fargoth]

Hi everybody, I have the GA-EP43-UD3L motherboard, with 1066 Mhz DDR2 and Core 2 Duo E7500. in my BIOS i set the CPU multiplier to x11, and the reported FSB is 266 Mhz. That's also what CPU-Z is showing. but PassMark shows my FSB as 246, and so the CPU is only running at 2646 according to PassMark... Who's right, and is there another good software which can check my real clock rates? (preferably from a boot CD, so that the OS won't be an issue).

[wonderwrench]

Are you using the latest build of version 7, build 1009? What OS are you using? 32/64 bit? I know PT 7 is rather new and the hardware info given is a big step up from PT 6 so I can see its still a work in progress. My Q6600 was being miss reported when PT 7 first came out but its right on the money now. I would guess getting PT to report correct info on most if not all systems is a lot of work. It most likely will never be correct on all systems, just too many variables. CPU-Z is not always correct and neither are mobo bios's. I'm sure one of the great guys from Passmark will chime in with more incite so stay tuned.

BTW its best to disable power management aka speed step before running PT. This is because if power management is declocking your CPU the hardware info and results can be wrong.

[Ian (PassMark)]

Most modern CPUs vary there operating frequency based on load algorithms (mostly modifying the multiplier based on the power state).

What has happened here is that the CPU measurement has been taken when the CPU is in a lower power state (ie. a lower multiplier) than the maximum. PerformanceTest attempts to max out the CPU for a short period to get raise the CPU to the highest power state (i.e. the maximum CPU frequency). In this case (for some reason) the CPU did not get to the maximum operating CPU frequency, hence in this case the CPU speed is not the maximum. The multiplier is determined to be 11. The FSB is then derived from the measured CPU speed and the multiplier, which is then wrong in this case.

If you would like us to investigate why the CPU did not get to maximum power state when measuring the CPU frequency, please let us know (our email address can be found here http://www.passmark.com/support/index.htm) and we will send you a private build to dump out some CPU information.

Regards,
Ian