[SOLVED] CPU reaching 100 degrees during prime95 test?

Toggle sidebar Toggle sidebar
G

Gharvar

Distinguished
Jan 17, 2015
38
0
18,540
I've been having a lot of issues with my PC getting whea uncorrectable errors while playing more demanding games lately. I've had the issue for years but it's gotten especially bad now. When my PC blue screen, it will repeatedly blue screen again while booting for up to 20 minutes but I noticed that if I close it for a while then it will just boot back up. Today I decided to test a bit and see if I could figure out the issue. I was using Corsair link to see the water cooling temp and core temp. During a prime95 test my cpu reached 90-100C degrees and crashed within a minute. So now I'm think the problem might be cooling?


I cleaned up my PC and reapplied thermal paste this afternoon which doesn't seem to have helped. Could my Corsair h80iGT be low on fluid or something after years of use? Should I try the stock cooler to see if the PC fares better?


CPU: i7 4790k

Mobo: Maximus hero 7

Water cooler: Corsair h80iGT
 
Solution
G
Thanks everyone for all the help. I finally got my parts in the mail. I changed my water cooler, some fans and got new thermal paste for the hell of it. I'm guessing PCpartpicker didn't recommend the h100x because it was a gigantic pain in the ass to install in my case but in the end it fits, just a tough install.


So far my PC had not bluescreened since I reinstalled the h80 but prime95 was still going to 100 degrees. Now with the H100x prime95 goes to 73 degrees max. So it was indeed the water cool being past it's expiration date.
Sort by date Sort by votes
Small FFT, all AVX options off. 10-15min run. That's it.

Blend runs EVERYTHING(smallest, small, large, etc), one at a time, for like 20mins, or something like that. So, 20mins of smallest, 20mins of small, 20mins of large and so on.
Plus, if I remember correctly, all 3 AVX* options are enabled by default... so if you didn't touch anything else, then you pounded the heck outta the cpu. [Some cpus don't support AVX-512]
 
Upvote 1 Downvote
Phaaze88 said:
Small FFT, all AVX options off. 10-15min run. That's it.

Blend runs EVERYTHING(smallest, small, large, etc), one at a time, for like 20mins, or something like that. So, 20mins of smallest, 20mins of small, 20mins of large and so on.
Plus, if I remember correctly, all 3 AVX* options are enabled by default... so if you didn't touch anything else, then you pounded the heck outta the cpu. [Some cpus don't support AVX-512]
Click to expand...
I did Small FFT with and without AVX and AVX2 it ran fine for 15-20 minutes but the temps still instantly climb to 91-100c. My cpu does not support AVX512 and it is disabled by default in Prime95. I tried a blend test without any AVX and it still bluescreened within minutes. I'm quite puzzled as to why it doing multiple bluescreens just before or after getting to desktop when it reboots, it's almost like they're progressing further, like the first 2-3 will be before the desktop even shows up and then it lets me get to the desktop but it will still bluescreen a few more times before everything is fully open.

I'll try a memtest86 over night and tomorrow switch back to the stock heatsink and if it helps at all. Hopefully this thing can be fixed. It's getting old but making a new computer these days seems like a pain in the ass with all the shortages.
 
Upvote 0 Downvote
Coolant temp. Rads aren't so much a heat exchanger as they are an energy exchanger. It takes a tremendous amount of wattage to raise coolant temps even 1°C because the coolant is continuously shedding any energy picked up from the cpu, into the rad to be dissipated. But it will eventually acclimate to the load, with Prime95, that usually takes close to half an hour.

Your cpu when under a high load like gaming is using power. The side affect of that use is wattage given off as heat. That wattage is absorbed by the coolant, goes through the rad, and dissipated. The coolant is just a medium of transfer, it's barely affected by the wattage it absorbs. The coolant going into the pump is generally only 1-2°C cooler than what is coming out.

So when you bluescreen, the coolant is already at high temps and the higher the coolant temp, the less effective it becomes at absorbing energy. Which means the cpu runs hotter, which gives off more heat etc in a viscous cycle of continuously getting worse.

And then you reboot before the coolant can 'cool off' because the pump wasn't running the coolant through the rad while the pc was off and is still overly warm.

The tubing in an aio is specifically designed to be as low permeation as it can be. Generally it takes 5-6 years before the coolant has lost enough oxygen and other smaller molecules that has broken down into gas form through permeation, which lowers coolant levels to the point where high pump demands mean you are pumping more gases than coolant.

In a nutshell, it sounds like the coolant is no longer effective and being a specific formula in a sealed system, you can't just add more. Time to retire the H80i.
 
  • Like
Reactions: MEMOFLEX
Upvote 1 Downvote
A little update. I did two tests with memtest86 overnight just to see if ram was ok and what not, everything came back good. I also tried to switch back to the stock heatsink that comes with the CPU with no success, it ran high in temp and crashed fast so I switched back to the H80iGT. For some reason after putting it back together my temps are a bit lower and I'm not crashing some games that would crash a few days prior. For example Bannerlords would be running the CPU up to 85-100 degrees constantly and bluescreen within 15 minutes. Right now the temps get in the 50-75 range. Prime95 still raise the temperature in the 90-100 range but it takes longer. I'm going to guess I maybe had a big air bubble and the fidgeting moved it something?


Still seems like the idea will be to get a new cooler. I'd welcome suggestions since corsair doesn't appear to be selling any H80 right or at least I can't find them on Newegg/Amazon. I was looking into a H100 but PCpartpicker seems to think a h100 wouldn't fit my Antem P100 case.
 
Upvote 0 Downvote
H80 is likely discontinued. H80i V2 still seems to be floating around, but it's only a matter of time before it disappears:
-Corsair US store appears to have it in stock.
-saw it from Amazon US, but the shipper was Amazon Global UK.
-Newegg has shipping from China($$$).


A quick image search of the Antec P100 - assuming you made a typo here - suggests that it does have room for 240mm AIOs up top.
 
Upvote 1 Downvote
Thanks everyone for all the help. I finally got my parts in the mail. I changed my water cooler, some fans and got new thermal paste for the hell of it. I'm guessing PCpartpicker didn't recommend the h100x because it was a gigantic pain in the ass to install in my case but in the end it fits, just a tough install.


So far my PC had not bluescreened since I reinstalled the h80 but prime95 was still going to 100 degrees. Now with the H100x prime95 goes to 73 degrees max. So it was indeed the water cool being past it's expiration date.
 
Last edited:
Upvote 0 Downvote
Solution
Stories like yours was the impetus for a couple of AIO vendors like Swiftec and Fractal Design to add standard water cooling fittings to their rads. Not only did that make a 280mm/360mm more usable as a whole system cooler for cpu and gpu (hybrids etc) but also allowed a very easy way to flush the old coolant and drop in some new coolant every couple of years. Takes 1 of the 2 possible failure points out of the loop, so to speak.
 
Upvote 1 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom