Question CPU is boiling hot, don't know why or how to fix it.

Sep 21, 2021
4
0
10
0
Recently I upgraded my GPU and RAM, and today I noticed the fan noise while playing some KSP. Looked at my i7 8700 temps, and to my horror, they were close to 90C. Booted BFV, and I even got a 99C high mark.

On idle, it stands around 46/60, roughly 20C higher than it's supposed to be, from what I read. I updated bios, changed thermal paste, turned MCE off, even turned turbo off, the temps remain roughly the same. I only saw a promising change when I changed paste, but the temps soon climbed again. Tomorrow I'm doing a clean windows install in case it helps. Whrn IU boot up, theres a huge temp spike, around the high 80s, then it settles on its now "usual" 50-60C. I read through the pinned thread, but honestly, I'm an idiot when it comes to this stuff. Any help will be greatly appreciated.

Here are my specs:

CPU is an I7, 8700, non-k, with it's stock cooler.
CPU is not OC's, currnetl at 3200MHz, when Turbo was on I saw it up to 4.6 MHZ iirc.
Max VCore seems to be 1.056V, while running Prime 95 for a few minutes.
Max VCore on startup reaches to 1.312 V, this seems suspicious.
Reading my temps from HWiNFO
My ram is 2x16 3200MHz Ballistix, set at 2666 MHz.
MB is an ASUS Z370 A Prime
Graphics is an RTX 3060ti
Ambient temp is around 24C.

On bios, Temps look perfect, around 30C.


Here are some snips from HWinfo, and CPUZ,

and some numbers that caught my eye after that temp spike during boot. Why are IA Cores Power and CPU package power triple than when idling?

XMP Profiles are off, and now that I remember, I've been seeing a few blue screens, more than before. Maybe 2 or 3 last month, when maybe I saw one every 3 or 4 months. Could the new RAM be the culprit?
 
Last edited:

nestlee65

Honorable
Feb 16, 2016
16
0
10,510
0
The same thing happened to me a few years ago. I replaced my cooler with a brand new one and it ran perfectly after that.
I also downloaded Core Temp for free and installed it. You might want to look into that.

Pierre
 
A Z370 mainboard can easily run a a non-K 8700 well beyond specs, quite close to 8700K power/clock speeds, which would indeed predictably and quite easily overwhelm the factory cooler....

If power limits and/or are boost duration are disabled, or MCE enabled, it might even attempt to run all cores at max turbo .(4600 MHz)..

(If HWMonitor shows all cores at 4600 MHz, then MCE is definitely enabled in your BIOS...; if those clock speeds are maintained all the way to 100+C , or even for more than 90-120 seconds, then boost duration limits are disabled, and likely power/TDP limits as well.)

Depending on your goals, you need a better cooler (that would be my choice!), or, can lower clock speeds to more moderate all-core boost speeds, which can be done easily with INtel's XTU application.
 
Last edited:
Sep 21, 2021
4
0
10
0
I'll probably invest in a cpu cooler and troubleshoot with that. MCE is definitely disabled in bios, is it showing at 4600 somewhere on the screenshots? Is it overriding my configuration for some reason?

Most of the advanced options on XTU appear greyed out for me.
 
Last edited:
Sep 21, 2021
4
0
10
0
Ok, tried some stuff, here are the results:
Undervolting Vcore resulted in a bunch of L0 cache results, so I let it come back to auto. So i decided to just turn everything down:
Turned off MCE, turned off intel's turbo mode, so now it's running at a stock 3200MHz, ram is at 2400.

And yet, idle temperatures while opening something like opera -granted, i have a monstrous amount of tabs open- are around 60C. I'm installing a coolermaster 120 liquid cooler today, but these temps seem to be way, way off, and I don't know how much difference a better cooling solution will make. Is there a way to diagnose where the problem is coming from? Is my motherboard overvolting without telling me about it?

current HWinfo capture, fwiw.

I'm worried this might have been going on for longer than I thought, I asked the guy who built my pc if the stock cooler was enough without OC, unknowing my motherboard would overclock it anyways. Learned a lesson here. In that regard, is there a way I can I tell if there's temperature damage done to the CPU? Is there some sort of diagnostic tool?

In any case, i'll give a last bump after installing the new cooling solution. Hopefully that was it.
 
Last edited:
Not to worry about temperature causing damage.
The cpu will monitor it's own temperature and will slow down or shut off to protect itself if it detects a dangerous temperature,
That point is 100c.
your numbers are not too bad if you are using the stock cooler.
What is the make/model of your case?
It needs good intake airflow to let your cooler do it's job.
I would consider a good tower type air cooler with a 120mm fan.
Noctua makes some good ones.

FWIW:
----------------how to mount the stock Intel cooler--------------

The stock Intel cooler can be tricky to install.
A poor installation will result in higher cpu temperatures.
If properly mounted, you should expect temperatures at idle to be 10-15c. over ambient.

To mount the Intel stock cooler properly, place the motherboard on top of the foam or cardboard backing that was packed with the motherboard.
The stock cooler will come with paste pre applied, it looks like three grey strips.
The 4 push pins should come in the proper position for installation, that is with the pins rotated in the opposite direction of the arrow,(clockwise)
and pulled up as far as they can go.
Take the time to play with the pushpin mechanism until you know how they work.
Orient the 4 pins so that they are exactly over the motherboard holes.
If one is out of place, you will damage the pins which are delicate.
Push down on a DIAGONAL pair of pins at the same time. Then the other pair.
When you push down on the top black pins, it expands the white plastic pins to fix the cooler in place.
If you do them one at a time, you will not get the cooler on straight.
Lastly, look at the back of the motherboard to verify that all 4 pins are equally through the motherboard, and that the cooler is on firmly.
This last step must be done, which is why the motherboard should be out of the case to do the job. Or you need a case with a opening that lets you see the pins.
It is possible to mount the cooler with the motherboard mounted in the case, but you can then never be certain that the push pins are inserted properly
unless you can verify that the pins are through the motherboard and locked.

If you should need to remove the cooler, first run the cpu to heat it up and soften the paste before shutting down and powering off the pc. That makes it easy to unstick the old cooler.
Turn the pins counter clockwise to unlock them.
You will need to clean off the old paste and reapply new if you ever take the cooler off.
Clean off old paste with alcohol and a lint free paper like a coffee filter.
Apply new paste sparingly. A small rice sized drop in the center will spread our under heat and pressure.

It is hard to use too little.
----------------------------------------------------------------------------------
 

faalin

Judicious
More then likely when you replaced the ram and GPU you probably bumped the CPU cooler. In doing so it broke the old past bond to the cpu, remove the cooler and clean the cpu and cooler and replace the thermal past.

Also what is the make and model of the PSU
 
Sep 21, 2021
4
0
10
0
Thanks for the replies, in order:

Case is a cooler master CM 690 II, it has good airflow, recently did a deep clean of all fans. Cables are managed through the back of the motherboard, so plenty of space. I reset the cooler yesterday, applied new thermal paste, and noticed that one of the plastic legs of the cooler was broken, probably during it's first installation. Otherwise, the cooler looked snug, and the paste was pretty well distributed.

I might have bumped the cooler, you are right. I will check temps once I install the new one.

My PSU is about 5 years old now, it's a 750W modular, pretty good quality one. Can't remember the maker now, will update this message when I open the case later today.

Do you think the PSU might be struggling because of it's age, causing temperature problems? I have a LOT of peripherals constantly plugged in, namely xbox controller and headphones receivers, a pioneer mixer that acts as a soundcard, a mic, 1 midi controller, an elgato capture card, a wacom, an external hdd, 2 internal HDDs and 2 SSDs, 2 moniotors and a projector hooked to the 3060ti. Do peripherals like those drain a lot of wattage? Online calculators tell me I'm more than fine, but I'm worried about the age.
 

ASK THE COMMUNITY