Built myself a new PC over the weekend and would like a little bit of feedback on load temperatures of the processor post OC
Case - Coolermaster HAF932
Cooler - Corsair H50 WC
Mobo - Asus P6X58D-Premium
CPU - i7 930 D0
Voltage currently @ 1.35v with HT enabled. BCLK @ 190 / Multiplier @ 21.0
After a ~90min session playing Bad Company 2, HWMonitor reported the peak temp of Core1 @ 81c - first of all, am I right in assuming this is too hot for the i7's? Secondly, to anyone with a H50 what sort of temperatures do you hit at 80%+ load?
It's highly likely that the thermal paste isn't particularly well applied - I stupidly hooked up the H50 at about 2am, totally knackered...
Just to add, there isn't a great deal of clearance at the back of the machine, 5inches perhaps. I'd also suggest that a good chunk of the air passed over the H50's radiator is hot air out the back of the 5970 exhaust.
Average room temperature is around the 25-30c mark. Is repositioning the desk/tower an idea?
The more outside air the H50 can pull in, the cooler it will be ofcourse.
Also if you don't have 2 fans on the Radiator, then another fan is going to make alot of difference. On this one, the difference was 7-9c under load.
I have a 920 at 1.231 vcore at 4ghz (same 190 bclk and 21x) and it stays at 68-73c under prime95 hours of testing. Vcore is going to make a huge difference in temps
I would try lowering vcore as low as possible and run tests. If prime95 crashes or stops, then up the vcore 1 more notch. Keep doing so until your stability is 100% through 8 hours of prime95
It HAS to be 8 hours... you can get a crash 4-5 hours in. 81c is too hot imo, but any real world application is not going to make it run that hot like prime95 will. Up to you, most people recommend keeping it below 75c under 30 minutes of prime95.
There is also kind of a "sweet spot" on vcore for i7's also..., On a 920 to go from 3.8ghz to 4.0ghz, I have to move the vcore from 1.224 to 1.231 which is just .007 more. But from 4.0 to 4.2, the vcore has to go from 1.231 to 1.263 which is .032 more. Big difference there. Once you get to the real high OC's, it starts asking for alot more vcore and vcore is the biggest part of creating higher temps.
Cheers for the feedback - will definitely bear your suggestions in mind.
For the time being I've changed my fans around and moved the H50 radiator & 120mm fan so that it now draws air from the top of the case. I've replaced the 120mm H50 fan previously in the upper rear slot with the 140mm fan shipped with the case, which now acts as an exhaust. Case temp. is relatively unchanged, though I've seen about a 15-20c drop in CPU idle/load temps. I've also reduced the vcore to 1.300 - which is stable thus far; primed for ~3hrs / 20 cycle IBT.
Reasoning behind the move was because of the poor clearance at the back of the case, the only possible air the H50 fan could have been drawing in was the hot air released by my PSU and more importantly, my 5970.....i've only got my own planning to blame for that...
The above IBT saw the cores peak @ 69/65/65/64 - am I right in thinking that the variance is temps is likely to do with poorly applied paste? Or can core temps vary quite so much?
And for the post change BC2 comparison, it now peaks at 61c after ~1hrs gameplay - looking good thus far.