q9300, once core runs 20c hotter

jeffline

[H]ard|Gawd
Joined
Mar 31, 2001
Messages
1,295
alright, so I just got a q9300, yesterday to replace my E6750

at idle according to HW monitor, and speedfan cores 0,1,3 all run within 5c(32c-37c) of each other but core 2 runs almost 20c(51c) hotter, WTF :confused::confused:
The heatsink is the intel stock for now, IT is seated properly and the fan is running @ 2200-2400 rpm, the case is an antec 900 with fans all set to medium speed, with an extra on the side panel
under load it does the same thing, core 02 almost hits 70c

what could be the cause of this ? its it just an error, or is core 02 bad ?
 
Could be a flakey sensor on the chip.

RMA it. I wouldn't settle for that.
 
Seems to me it might also be a matter of the thermal interface material.

Did you just go with the pasty stuff that came on the Intel cooler,
or did you use a Real thermal interface material, like AS5 or MX-2?

Also, you might want to double- and triple-check with alternate
utilities, perhaps PCWizard from CPUID.com, or RealTemp.
 
Most cores are a few degrees off from each other but that seems a little out of the "acceptable" range to me.
 
If you could rma it, go ahead, otherwise, i would buy a different cooler and see if that fixes it, i had a similar problem with my Q9550 when using the stock cooler when i had no tim left.
 
45nm Intel chips are KNOWN to have flaky temp sensors. My E8400 certainly does, reads about 10C higher then it should(I've taking a temp sensor to it). Either RMA or ignore it.
 
Yup, a friend of mine has the same issue with his Intel system, with all four core's temperature probes giving nonsense readings.

Best test is to just feel whether the HSF can still be touched without scorching your fingers ;)
 
Yup, a friend of mine has the same issue with his Intel system, with all four core's temperature probes giving nonsense readings.

Best test is to just feel whether the HSF can still be touched without scorching your fingers ;)

Hardly a good way to read it XD a lot of heat sinks can cool themselves better then the chip, rofl. But reguardless, I'll be honest with you. Do not worry about it. It's like this, if the chip gets too hot and risks damage, it will thermal throttle. If it continues to get too hot after the throttle, it will shut off, don't sweat it, I'd imagine your core is the same as the others within 2-5 degrees, relax and enjoy it.
 
The problem with thermal throttling is that he isn't getting the performance he paid for. Thermal throttle will slow down the chip which means he's out at least one core.

I would test the chip, maybe with Prime95 and see how each core handles the load. If you notice that the prime 95 set to that core fails a lot easier than the others I would RMA it.
 
its probably the thermal sensors. do they reach a point where they will not go any lower, even if you lower your vcore? i have had many, many 45nm chips, dually and quad, and many of them had a broken sensor or 2. they usually were broke in the downward dirwction, ie they would measure temp accurately above say 30c, but would stick at 30c going down.
 
Back
Top