CPU stupidly hot, suggestions....

[roachcoach]

I'll have a check. I'm planning on logging the temps in BF3 tonight once I get it running - as I say, its only under extreme duress I'm seeing this sort of thing. Provided it's ok there, I'll pull it apart at a more convenient time and reassess.

I was always planning on an aftermarket fan at some stage for OCing anyway, maybe as good a reason as the next to bring that purchase forward.

I've never had any issues in the past, its been solid as a rock and this isn't my first PC built, which is part of the reason I was so shocked that the core (indeed cpu per PC probe) was hitting those kind of temps. However that needs weighed against the the fact there's not much in the way of games right now that clobber the CPU to the roof, as I say, 3dmark11 got it nowhere near max temps and that's probably more representative of gaming which is the 'typical' peak pressure I'll apply to it.

That and I was pretty damned sure I had a nice powerful cooling setup, it's a nasty shock to see you're mistaken when you put the CPU through the mill.

@PeterB: To be fair to PC probe, its reporting CPU temp, not core temps. realtemp was first drawn to my attention from the [h]ardOCP forums and a lot of folks over there know what they're about (as best as I can tell). Even high 80's feels a bit high to be honest, just not as as 100

[roachcoach]

Update: BF3 isn't stressing it, so I'll hang fire for a bit on panic stations

[Noxvayl]

I must say I've not come across software with such disparity in sensor readings before. I use HWMonitor, AIDA64 and check them against BIOS readings; I find them to be similar.

If I found those results on my machine I'd do the following:

• Open the case and feel the heatsink when running the stress test.
• I'd then check to make sure none of the fans in the case have stopped working(I've had that happen a couple times) and double check they are blowing in the correct direction.
• If it's still hot I will proceed with a thorough clean of the case, and everything in it, and do the tests again.
• If that doesn't reduce temps satisfactory then I'd re-seat the heatsink.

To be honest that is over cautious but I like to make sure everything is in perfect working order for my PC; I'm someone who laps his heatsinks to ensure optimal heat conduction. I'd imagine your PC is pretty safe considering it won't be stressed anywhere near prime95 levels.

[Zybros]

Monitor the temperature of the CPU idle times & load times over a period of a few days to see whether it's a one off or a continuing issue. I've got a AMD Phenom 9500 and it runs nicely at around 31C with a limited edition scythe Ninja Full Copper CPU Cooler, havnt seen any for sale in ages though but you could try ebay.

[roachcoach]

I had to put a new sound card in at the weekend (thanks punkbuster >.<) and I took the opportunity to check the CPU/fan, and the TIM looked a bit....meh. I'll get some more today and get it reseated etc.

Fans are all fine

[shaithis]

If your going to re-apply, let me re-affirm that when applying it: 'less is more'. Your trying to fill the micro-gaps on the chips surface, not adding a sandwich filling

[scaryjim]

As everyone else has said, the first thing you need to do is try a multitude of monitoring software to get a "consensus" temperature. If you haven't done so already, grab a cpoy of coretemp as this reads from the sensor on the CPU itself. Compare this to the speedfan, realtemp and PC Probe temps (I've always found PC Probe, in various iterations, to be pretty reliable...).

The second thing to do is - stop comparing a 130W desktop CPU to a 35W mobile CPU. Your laptop doesn't have a better cooling system than your desktop, it was about 100W less heat output to deal with!

Until you've got a confirmed temperature for your CPU - which I don't think you have yet? - you can't really tell if there's a problem or not. Reseating the CPU cooler is unlikely to cause you any problems, but it may not be necessary, so you have to ask yourself if you mind going through the hassle of doing it "just in case". Otherwise, grab *more* monitor software and try to figure out what temps your CPU is actually running at...

[roachcoach]

@shaithis: Yes, research has been done long ago


@scaryjim - well the thing is, unless both are overreading horribly (unlikely) it's still (far) too hot for my tastes (both software are throwing warnings). I'll get it all sorted out - reseating it might be a hassle but I'd rather be safe than sorry right now, particularly since this all started because I was looking at OCing it.


There are differentials in the software used but they are relatively consistent gaps. As I mentioned before, one is showing core temps, the other cpu temps, that could account for the differences but it still doesn't make me happy about the temperature the thing is running at. High 80's are still a problem for me, given it's a 100% stock chip.

[roachcoach]

I grabbed a new cooler whilst I was out, upshot is Prime95 torture tests are now 40 degrees less in realtemp, PC probe values will follow via edit later once I load it and test that.

Aftermarkets are good, but I reckon a 40 degree drop would suggest something was very much awry with the original configuration.

Edit: PC Probe is following a similar gap to previously, again a similar drop to prior performance.

[Noxvayl]

What cooler did you get and what cooler did it replace?

[roachcoach]

Cooler master hyper 212 plus, replaced stock

[roachcoach]

Also, speedfan and realtemp are the same readouts. Speedfan CPU temp matches the PC probe temp.

Just me being a muppet and not noticing the original distinction between CORE temp and CPU temp

[peterb]

Glad you got to the bottom of that one!

Now the interesting question is "What are the temps if you refit the stock cooler?"

[roachcoach]

There's not a cats chance in hell I'm doing that again unless something breaks!

This case is big but its still a very busy place and full on mobo extraction is something I'm not doing for fun, far too time consuming!

[peterb]

There's not a cats chance in hell I'm doing that again unless something breaks!

snip--->

Now why am I not surprised?

[roachcoach]

Well tbh if I didn't need to muck about with back-plates and all that nonsense, i.e. if I could do it 'in the case' as it were, it would be worth the test...but I would...soooooo....y'know