[SOLVED] Is 80-90°c safe when stress testing an overclocked i7-7700k (4.8Ghz)?

Jun 26, 2020
11
1
15
0
I recently got a PC with a 7700k for a pretty cheap price and started overclocking recently. My motherboard (gigabyte-z270-hd3) has inbuilt overclocking profiles for the i7-7700k. I've been using the 4.6ghz profile for a few days and when stress testing it only reached 84° at max but stayed around the 70-77° mark for most of the test. My motherboard also has a 4.8Ghz profile which i tried out, during the stress-test it peaked at 91° but stayed around 80-88° for majority of the test. My cooler is the ThermalTake Contac 12 Silent (Without the voltage/rpm limiter).

I want to know if these temperatures are fine or if I should switch back to using 4.6Ghz. I'm probably going to switch back to 4.6 but would appreciate some feedback or advice
 
Last edited:

Darkbreeze

Retired Mod
No, it's not. Anything over 80°C is undesirable. Anything over 85°C is getting specifically into "this is where stuff starts going wrong" territory in terms of premature thermal damage, electromigration and VT shift.

READ this, which will tell you practically everything you could want to know about thermal accountability on Intel processors.



Also, what "test" are you running to determine your maximum temperature? You need to be aware that "stress" tests and "thermal" tests are NOT the same thing. Stress tests generally indicate stability testing, which involves a variety of variable, fluctuating work loads. Thermal testing requires a utility with a steady state workload that results in the CPU being at or near TDP. Generally speaking there are two pretty acceptable thermal steady state work load utilities. Prime95 Small FFT with all AVX instructions disabled, or OCCT small data set, also with AVX either disabled or offset. I, and most others, prefer Prime95 because it is marginally closer to full TDP than OCCT small data set and more easily configurable.
 
Reactions: MirageTPX

Darkbreeze

Retired Mod
No, it's not. Anything over 80°C is undesirable. Anything over 85°C is getting specifically into "this is where stuff starts going wrong" territory in terms of premature thermal damage, electromigration and VT shift.

READ this, which will tell you practically everything you could want to know about thermal accountability on Intel processors.



Also, what "test" are you running to determine your maximum temperature? You need to be aware that "stress" tests and "thermal" tests are NOT the same thing. Stress tests generally indicate stability testing, which involves a variety of variable, fluctuating work loads. Thermal testing requires a utility with a steady state workload that results in the CPU being at or near TDP. Generally speaking there are two pretty acceptable thermal steady state work load utilities. Prime95 Small FFT with all AVX instructions disabled, or OCCT small data set, also with AVX either disabled or offset. I, and most others, prefer Prime95 because it is marginally closer to full TDP than OCCT small data set and more easily configurable.
 
Reactions: MirageTPX

Darkbreeze

Retired Mod
If you were hitting those temps at 4.6Ghz, then I'd try reducing voltage slightly and then stability test it. If you can't hit stable 4.6Ghz with a low enough voltage to stay below 80°C, then reduce your OC to 4.5Ghz and try that with a reduced voltage. Mostly, you need a more capable cooler, and depending on your case and case fan configuration, you might need some improvements there as well.
 

ASK THE COMMUNITY