• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Critical CPU temperature exceeded and nothing bad happens

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Bucic

Registered
Joined
Jul 19, 2014
Location
Poland
It seems that maximum operating temperature for my CPU is 62 deg C.
http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series FX-6300.html

(CPUID software used for monitoring)

Recently I've checked my over-the-top OC config with Prime95. Temperatures went to 75 deg C quickly and kept rising. I went back to BIOS to make sure 'Hardware Thermal Management' is enabled and went ahead with P95, keeping my finger on the power button. I terminated the testing at 85 deg C (after a minute or so) as at this point I think I need some help understanding this. I've repeated the test once more (up to 85 deg C) making sure that:
- voltage is not getting capped (HW Monitor) ; it didn't
- CPU clock is not getting capped (CPU-Z) ; it didn't

Do you guys have any possible explanations to all this?
 
a screen shot is worth a thousand words. Post a ss of Hwmonitor so we see what you see.
 
Last edited:
the post in the link shows the 6300 at a peak temp of 40c.
tempin2 is the socket temp and should be held below 72c.
the temp line of amd 6300 is the core temp and should have a max of 62c.
 
TMPIN2 is my only temperature exceeding 40 deg C.
This is the socket temperature. Some fans on the VRM and behind the motherboard can lower this a bit, with your board the VRM are going to get taxed and heat up. I'll post a couple of pics of what can aid this. It's also imperative that there be "plenty" of fresh air moving through the case. My set up had ~ 300cfm moving through when everything was cranked up. The actual core temperatures are listed as the "package" temp since the FX doesn't monitor individual cores.

Capturefan.PNG

vrm fans.jpg
 
the post in the link shows the 6300 at a peak temp of 40c.
tempin2 is the socket temp and should be held below 72c.
the temp line of amd 6300 is the core temp and should have a max of 62c.
This would mean I bought a new cooler having a 20 deg C margin :)
Can you provide a source for the 72 being max socket temp?
What do you mean by 'the temp line'?

Fishy stuff:
- at low loads my package temp is reported as ~20 deg C. I've seen it at 12 at idle! My case temp (bottom) is 26.
- is it even realistic for my linked 4.5 GHz at 1.35V to stay below 42 deg C with my cheapo cooler?
 
The FX sensor doesn't keep accurate temps below 40-45c. That's why you'll see 12c in a 25c room which is not physically possible unless you're using some exotic cooling.
 
Thank you all guys! This changes so much!
I get that really (actually) exceeding specified maximums DOES affect component wear significantly /kills it on the spot. OK, all clear.

@Johan
Thank you. Case closed then.
 
I have this board and did some testing on it a few months ago.

TMPIN0 is tiny resistor in the lower left corner of the board, above the ITE chip. It measures ambient air temp/general motherboard temp.

TMPIN1 is a unkown/error in temp calculation between CPU package and TMPIN2. (It may be socket temp)

TPMIN2 is a constant 15 degree offset added to the cpu package. This is done to give realistic idle temps, since the cpu package is not accurate below 40c.

CPU package is the only real temp we have available. AMD overdrive lists the 6300 as 70c max termal margin and 80c +/- TJmax which is major throttle down/shutdown. Most user's chips do not play well above 60c.


Also the VRMs on this board have a heatsink. You can point a fan towards the heatsink to help with overclocks.
Or you can redo the themal pads under the heatsink and put real thermal paste.

I have the same board and chip at 4.6 and ram at 2136. You can go higher cpu clocks but need a high end cooler.
 
TPMIN2 is a constant 15 degree offset added to the cpu package. This is done to give realistic idle temps, since the cpu package is not accurate below 40c.

CPU package is the only real temp we have available.
Thanks for chiming in. So in your opinion I should disregard the TMPIN2 or what caddi daddi said remains applicable (TMPIN2=socket temp<72 degC)?

AMD overdrive lists the 6300 as 70c max termal margin and 80c +/- TJmax which is major throttle down/shutdown. Most user's chips do not play well above 60c.
I'm not sure I follow. Which HWMonitor temps correspond to these values?
 
Sorry I wrote that a little quick.

In HWmonitor the one you want to pay attention to is CPU package. You want that to be below 60c. You can go higher technically but it's not recommended.

Our board doesn't have an actual socket temp reading so 72c doesn't apply. But you can still use 72 or 75c as you max temp for TPMIN2. TMPIN2 is just the cpu plus 15 degrees at all times. You can see this for yourself if you get a program to log the temps into a graph output.
 
Sorry I wrote that a little quick.

In HWmonitor the one you want to pay attention to is CPU package. You want that to be below 60c. You can go higher technically but it's not recommended.

Our board doesn't have an actual socket temp reading so 72c doesn't apply. But you can still use 72 or 75c as you max temp for TPMIN2. TMPIN2 is just the cpu plus 15 degrees at all times. You can see this for yourself if you get a program to log the temps into a graph output.

Good info.

I've just tested with a ~4000 RPM GPU fan blowing at the back of the socket. It didn't make a difference in my setup. Neither at full nor at half the speed. No problem though. I still have plenty of t2 margin.

EDIT
OK, I need more sleep :/ Of course it didn't make a difference. It's a dummy indicator! Core temp + 15 deg C as manid said.
EDIT

One important consideration though (semi-OT):
How we should approach a potential socket temp drop should depend on the location of the actual temperature probe location. If we blow directly at the probe, I don't think we should take the drop 1 to 1.


Anyway, I'm all set. Thank you all, guys!
 
Last edited:
Back