CPU at 100C underload yet PC always fine? Am I in trouble

[Long Trinh]

Hi all,

After a change of GPU I decided to run some heat monitoring to see if everything was correct.

So the GPU has absolutely normal temp, but I realize something is off.

My CPU temperature underload, playing games such as Total War Warhammer, The Witcher 3 or Doom, would always be in the 95-100C.

This seems really off to me because I've been using this rig for 2 years (built myself) and it has always run very well with heavy load. I never had a blue screen of death or auto shutdown or anything. The only thing that happens is slower fps during large battles, but nothing else.

This is very strange because 100C for hours after hours would melt every CPU and motherboard, yet I mostly play these games for hours on end during the weekend.

Anyway I got a new CPU heatsink (I only use the thermal paste that comes with it). No changes at all to CPU temperature...

Idle is at 50C (BIOS reading), under load is 100C (CPU Z and MSI afterburner reading)

Is it actually a problem?

On one hand I feel like if the system has always run fine for 2 years no reason to fret out, but on the other hand...

Intel Core i5-4590 (3.3 GHz)
MSI H97M-G43
Geforce 1060

Thanks

 

[GhislainG]

"This is very strange because 100C for hours after hours would melt every CPU and motherboard" The worst that happens to the CPU is throttling and you can easily see it if it does if you monitor the CPU with Intel XTU. What is the new cooler? Is you case airflow good enough?

 

[Long Trinh]

Ok so I run some stress test with Intel XTU like you suggested. The only throttling that occurs is thermal throttling, but it only appears a dozen times every 5min, and when it does, only for 1-2 sec. Is that bad? is that why my fps goes down during large battle?

Here is the new cooler I use. Very popular among gamers in paris here:

http://www.ldlc.com/fiche/PB00216376.html?gclid=EAIaIQobChMI3e3cvOXC1QIVrhbTCh16Ug8-EAQYAyABEgICO_D_BwE

Got a big cooler master case with good airflow and I dust it off regularly. 1 fan front for air in, CPU fan for air out and up through the case's top, and back fan for air out in classic set up.

I dont think airflow is a problem since GPU barely reached 70C even during heavy gaming.

I dont see what else I can do. Maybe go out and grab an Arctic Silver 5 and try reapplying it?

 

[GhislainG]

Intel XTU should be used for monitoring only; its stress test is useless. If your CPU throttles when using that stress test, then there definitely is a cooling issue. Throttling directly affects FPS because the CPU frequency drops. I read the comments at the link that you provided; the first one says "En termes de performances c'est vraiment puissant. Mon i5 4460 ne dépasse jamais les 60 degrés en burn test et reste en dessous de 45 en utilisation normale." I would expect your CPU to run a bit warmer, but it shoudn't reach 100°C unless your ambient temperature is higher than normal. Does it make a difference if you leave the case open?

 

[Long Trinh]

No Ive been leaving the case open the last hour, with my room fan blowing at it and theres really no difference. Ambient temperature over here is around 20C, but CPU temp at idle is around 50C. Way too high for idle temp and 100C at gaming load is strange.

Plus there's really not that much difference between the stock cooler and the new cooler, despite the fact that the new heatsink and fan is way more powerful.

Left XTU on last hour gaming Total War (most CPU heavy game right now) and it shows CPU frequency around 3.1-3.3 (mine CPU is 3.3GHZ)

So could it be considered a frequency drop or is it normal?

Is there anything else I can do other than applying a better thermal paste?

Thanks,

 

[GhislainG]

Does it show throttling? The frequency drops when the CPU isn't that busy, but it should never throttle, even at full load.

 

[Long Trinh]

Ok so after reading a while it might seems the problem is with the infernal intel push pin installation. Maybe one of the pin doesnt push all the way through.

I'll try reapplying the thermal paste, remove the motherboard and make sure the heat sink is 100% snugged in.

Thanks for your help, will update if I find a solution

 

[GhislainG]

Ideally you can check from the back to ensure all 4 pins are fully locked in.

 

[Long Trinh]

Ok so after an HOUR wrestling with the pin I finally found out why. It seems like the heatsink was entirely secure, but the 4th pin wasn't 100% engaged. At risk of breaking the mobo in half I finally managed to push the sucker down. Temp went to 20C at idle and max 50 underload holy shit! A bit of Arctic silver 5 helped too.

Intel push-pin heatsink gotta be the biggest engineering fail of the century, cause this solution was suggested from a lot of other people.

Thanks for the help mate!

 

[GhislainG]

Glad you got it resolved.