Showing results for 
Search instead for 
Did you mean: 


Adept I

Running Cinebench on 3600

I have installed a Noctua 12s and I have ran cinebench and my max temp reached 81 degrees.

Does this sound ok with the stress test?

When I had the stock cooler on previously, I was hitting 95degrees when exporting from Lightroom. I can't imagine lightroom being more demanding on the CPU than the cinebench r20 test?


16 Replies

Did you remember to apply thermal paste?

Seems way to hot


I did, I added about a rice grain worth so more than enough. 

Worth mentioning that I very rarely touch the 80 degree, this was during a very heavy stress test using cinebench.

Before I installed my new CPU Cooler, I was getting 10+ degrees higher when using programmes like Lightroom so it has made a difference.

On Idle I am averaging between 35-40 degrees which seems normal to me?


An 80 degree max temperature is perfectly fine.

Idle of 35-40 also fine.


While within thermal limits according to the spec I think it is unusually hot too. 

I have had 2 3600s and neither get hot lock that. In fact I have never seen either under full load over 73. 

I run a CM 212 evo. 

I would make sure that you are making good contact with the surface of the cooler and the processor surface.

Make sure you don't see any unevenness in the heat pipes. 

Also make sure you are using good thermal past. I would use a dab about the size a green pea. Lots of opinions out there on what the right way to do this. I always pre spread it out and put a very thin layer on the cooler base too. 

I have never had an issue.

If that doesn't help and you are not having the performance crippled by throttling then you within spec. If you think you are losing performance. Talk to AMD support for advice and if need be start the RMA process.     For warranty: 


Thank you for your reply.

I think there are multiple factors that could influence the CPU temp ( Case, Cooler, Paste, Ambient Temp etc).

I put Artic MX4 paste on and made sure I put at least a pea size amount on the CPU and fairly confident this isn't my issue. 

I am seeing alot of people saying that as long as you don't reach 95 degrees (although the lower the better of course), then everything is fine. I ran another cinebench after changing my power settings and I got a maximum 79degrees which I am happy considering the factors (25 degrees ambient being one).

What do you class as full load? Surely Cinebench test will put it under full capacity stress?

100% processor load about 10 minutes. CPUz  built in benchmark stress test is usually sufficient to get the processor as hot as will ever get. 


I have just done a test using CPUz included stress test and I got a maximum of 84 degrees over 10 minutes.

It is still too high for my liking but I will never be using 100% of my CPU for 10 minutes ( not gaming) straight.

The only thng I can think of is have I put enough thermal paste on as everything else is working as normal.

Although having said all this, I think realistically, I may have quick spikes to the hgh temps but it will never stay more than a couple of seconds at the most so I should be fine. 

I also don't think the heat does me any favours so I am going to wait and do more tests in a cooler environment and if the issue starts causing me issues such as blue screen or shut down, I will reconsider my approach and get in touch with AMD. 

Thanks Bud!

Have you seen how opening your case affects the temps. Maybe you just need more fans or need to optimize things a bit for airflow. 


Prime95 "small FFTs" is by far the best test for heat.  Everything else is just a moderate to high load, Even CPU-Z.  If you are worried about your setup and if something is not working as it should with regards to temps...  just run Prime95 small FFTs and really push it to the limit.  The reason why you want to do this is that you never really know what kind of load an app or a game or an encode or anything is going to put on your CPU...  Not all 100% loads are created equally.

If the small FFTs torture test pushes your CPU past 95°C then something should be done to lower it...

For me, I have PBO enabled which pushes up the temps far too high for my liking even with my iCUE H100i RGB PRO XT..  so I found a nice under-volt which is rock-solid stable with PBO.  At first it seemed an offset of -0.14375V was good, but after a few random crashes I settled on -0.13125V.

Now I stay just over 80°C in the P95 small FFTs...  I only post the measures I have taken to illustrate that there are many different steps one can take to lower temps.




I have ran the prime95 tool for just over 15 minutes and the max I am getting is 82 but it fluxuates between 80-81 degrees.

I did the 'small ffts(testsL1/L2/L3 caches,maximum power/heat/CPU stress test)' .

I have stopped it after 15 minutes as the temps don't seem to be moving. 

How long do you leave it going for and do you think after 15 minutes it is a fair test?


There are plenty of utilities that will push the limits of a process. Everyone seems to have a favorite. Exactly as you said if CPUz causes over heating it shows you have an issue. You don't have to run prime 95 all night to know. 

I personally like CPUz as it is one utility that gives me a lot of information and I don't have to install many others. There are plenty of good utilities for that job out there for sure. 


With an air cooler 15 minutes should be plenty of time to find the max temp...  unless of course you have a case with bad airflow, then it may take a longer to heat up the air inside the case.  But if your max is under 85°C for this torture test I would put your fears to rest...




In my experience, having a good air flow in and around the case may help with CPU cooling and all temperatures. If introducing fresh air and extracting hot air are not in equilibrate, you can stuck hot air flow around the CPU cooler like in oven.

To see them you can experiment with front fan introducing fresh air and rear fan extracting hot air:

  • closed case, no fan running on front and no fan running on rear of the case : most hot
  • closed case, front fan running high speed and no rear fan running : heat down a bit
  • closed case, no front fan running and rear fan at high speed : heat down a bit more
  • closed case, front and rear fan running at same high speed : heat down significantly, best values
  • open case and do the same and compare.

Air flow vs speed vs noise depend on fan manufacturer.

Position of the fans in the case also may help, specially the front fan in the middle of the height or at the bottom to give fresh air to video card and CPU.



I did another test after making a couple of minor adjustments nd I don't break 76 after 15 minutes of stress testing prime95. 

This is 8 degrees cooler than my previous one.

It turns out I had my front fans on the wrong way so no air was going in to the case, it was being pushed out.

That will do it. This is why I had suggested running with the case open. If the temps drop by a lot you know you have a flow issue. Your temps are in line with mine now. Good job figuring it out!

Journeyman III

I had 85° C on the stock spire cooler. I replaced it with a corsair h100i Platinum liquid cooler and it came to 70 on the 10 min test.