Originally Posted by EarlGrey
I have a question regarding throttling.
I've noticed that when I'm using a program like intel burn test to really stress my cpu, realtemp 3.70 indicates my multi and cpu speed fluctuate wildly, i've seen it go as low as 500mhz even though the cpu is under full load the multiplier is constantly changing. At the same time I'm also running cpu-z and that program tells me my Q9650 is continually running at a solid 3825 mhz with only very minor fluctuations of 0.20ish mhz.
EIST and C1E are both disabled, so I'm wondering if there's any thermal throttling going on.
When using intel burn test, max ram, my temps are about 73c. With the TJmax in realtemp set to 100.
I think that's fairly high, but luckily during gaming my temps rarely if ever exceed 60c.
I've tried a third program called throttlewatch for but some reason it keeps giving me a driver error.
I currently have an E5300, which functionally is a crippled E8xxx C2D, which in turn is half of what you have. What they all have in common is that they all use the wolfdale die and thus react more or less the same with some exceptions due to the extra die in the package.
Anways, it sounds like you have some kind of BIOS-based thermal throttling or something going on, because when my E5300 is put through an IBT run, it can (and has) under extreme conditions reach 95C+. Even at that temperature the clocks remain stable at their set speeds. Dig through your bios and disable all thermal throttling and safety items, and see if that takes care of the wild clock swings.
Also, download and run CrystalCPUID, and see what the real-time clock window says (F4 when the program is up and running), and if all else fails try using it to force a specific multiplier. It just might work.