i5 4690k temps too high with a coolermaster hyper 212 EVO.

Gwad

Honorable
Feb 8, 2014
3
0
10,510
I decided to run core temp for the hell of of the other day and was horrified to see that my cpu was idling at 40c, way up from the 27c when I originally installed the cpu cooler months ago. So, naturally I remounted the cpu cooler. But today while running prime 95 I hit 70c. This seems way to high for a chip that's not overclocked. I am using the stock thermal paste that comes with the hyper 212 EVO, but even so its seems way to high. I was hitting highs of around 60c while gaming. I don't know what else to try, any suggestions?
 
Solution
The problem with p95 after 26.6 for Haswell+ cpus lies in the avx instruction sets among others. While it overheats the cpu far beyond normal temps, what it doesn't do is stress the cpu. If a core had a value of 1-10, avx being just 1, you are driving that 1 to extreme temps, while not using the other 9, which is good for high temps but lousy for stress. P95 v26.6 has less emphasis on avx and more emphasis on the other 9, equating to a better spread of cpu usage using more instruction sets of various types, that match up with a variety of programs. It's a more even stress test, even if it doesn't produce the spike heat from a single instruction.

The other components mentioned are motherboard components, and the air that bleeds out the...

brik94

Reputable
Nov 24, 2014
186
0
4,710


Yea, time to upgrade that cooler.
 

Captainzach

Reputable
Dec 22, 2014
213
0
4,710


The stock thermal paste is kind of bad in quality, that is why when I had the 212 I just purchased Noctua Theramal Paste, and a quick note is to not put too much paste on and have good airflow like atmos929 said
 

Gwad

Honorable
Feb 8, 2014
3
0
10,510

There is a bit of dust on the fins. Though it is a very thin layer. There is one outtake fan at the top of the case, and one intake at the front (It's a NZXT Source 210 Elite.) The fan on the 212 is blowing through the fins at the outtake fan as shown here.
uB5TQLv.jpg
 
I was using arctic silver instead of the cooler master paste with my 4690k and EVO 212. During the maximum heat test in prime95 mine was also hitting 70-72 C. Idled 35 C. This is with a push and pull fan, both fans Cougar 120mm fans.

I found lapping the bottom of the heatsink was worth 5 C at idle and under load. Didn't take that long and made a considerable difference. My last build was a 3570k, and I just think this generation chips run hotter.

evo212.jpg
 

atmos929

Distinguished
Apr 21, 2010
517
0
19,160
how about the cooler fan?... monitor temp and CPU fan RPM when stressing the CPU... the hyper 212 evo should be reaching almost 2000 RPMs in the worst case scenario, that is full throttle.

I would use prime96 for torture test -> blend test... this makes sure to make each core go 100% load

use hwinfo to check temps and RPMs
 

Smudgy

Reputable
Dec 14, 2014
313
0
4,860
no 70c is fine running prime 95 my 4770k hit 72c here's something to ease your mind though 40c is high at idle
http://www.eteknix.com/intel-core-i7-4790k-devils-canyon-processor-review/9/
your temps will be in the same range as the ones i showed for 4790
 

Gwad

Honorable
Feb 8, 2014
3
0
10,510



Alright so running blend test i'm hitting 53c as a high temp, and my fan seems to be running at about 1600 rpm. I did just remount this cooler last night, so is this just the thermal paste settling?
 

Captainzach

Reputable
Dec 22, 2014
213
0
4,710


If you didn't run a test after you put the paste on then yes
 

Karadjgne

Titan
Ambassador
You have a 4690k. With that generation cpu and after, you shouldn't run any version of prime95 after 26.6. The newer versions use extensive avx instruction sets that give totally unreliable and unwarranted temps as no other programs use avx to that degree. DL v26.6 and try your temps again.

Also to consider is fan curves. With the stock cooler it had an entirely different curve than what the hyper212 needs, so will run odd speeds. You will need to adjust either duty cycles or fan speed/temp settings in either software or bios. Stock fans can reach speeds of 4000rpm, the fans on the Hyper212 won't, so a duty cycle starting at 40% at idle will have some temp differences.

Lastly, running push/pull on a hyper212 EVO does nothing. It's a waste. You'll maybe see a temp drop of @ 2° all for the sake of increased noise, more draw on the fan header, and heating of the surrounding components that would normally be cooled by bleed air from the sides of the cooler that are now redirected by the draw of the second fan. Totally unproductive.
 


Seems to me that makes it the ultimate stress test? It will make the CPU hotter than any other program, isn't that the goal?



I wouldn't say that's a waste. It's a drop in temperature, and that's the goal. In my case, I'm using quality, slow moving fan, and getting the drop in temperature, and no noise to speak of. For most uses my fans are only at 600 RPM or so. There is plenty of air flow in my case so I really don't care about the other components.

I just don't understand this thought process. Many people overclock CPUs for a relatively small gain. Why not use a second fan for a small gain there too? Fans, even nice ones really aren't exspensive.

 

Karadjgne

Titan
Ambassador
The problem with p95 after 26.6 for Haswell+ cpus lies in the avx instruction sets among others. While it overheats the cpu far beyond normal temps, what it doesn't do is stress the cpu. If a core had a value of 1-10, avx being just 1, you are driving that 1 to extreme temps, while not using the other 9, which is good for high temps but lousy for stress. P95 v26.6 has less emphasis on avx and more emphasis on the other 9, equating to a better spread of cpu usage using more instruction sets of various types, that match up with a variety of programs. It's a more even stress test, even if it doesn't produce the spike heat from a single instruction.

The other components mentioned are motherboard components, and the air that bleeds out the side of the tower is then drawn by the exhaust fan, creating airflow over the upper and lower sections of the motherboard surrounding the tower. This includes multiple caps, diodes and other temp sensitive components. You may have great airflow in your case, but with the placement of fans, this doesn't automatically mean you have great airflow in all areas of the case.
 
Solution