Stock 6700k hitting 94 Degrees on Prime95

Hey guys,

I just finished my first PC build for the second time (the first time my usb 3.0 header was damaged). I have an Asus z170 pro gaming aura mobo with a 6700k cpu.

My stock cpu is hitting 94 degress at 2.944 volts within seconds of starting Prime95's blend test. At idle it is at around 33-36 degrees at 1.440 volts with spikes every few seconds up to around 2.3 volts.
I know this is not good at all, so I will be dismantling my pc for a second time to return the processor.
Will these high cpu temps and voltages have harmed the mobo as well? I believe I am still within the store warranty period for the mobo and can swap it for a new one.

My main question though is: Why is this happening, and what can I do to fix it if it happens again with my new cpu and mobo?

Like I said, this is my first build, so I am new to a lot of this. It is extremely frustrating to have to spend so much of my time and energy taking apart and putting back together this pc multiple times
Oh, and all of my BIOS and mobo drivers are fully up to date. I'm not sure if I have to uninstall/reinstall the drivers after putting the new mobo and cpu in.

Thanks for your help.

Please provide full specs including case, cooler, etc that you're running. The 6700k doesn't come with a cooler so I assume you have an aftermarket one? What is it? Your voltage is also quite high. 1.431v seems very high to me.

Full specs are as follows:
- Fractal Define R5 case w/ 2 Fractal 140mm GP-14 intake fans on front and 1 140mm GP-14 exhaust fan at rear
- Hyper 212 Evo CPU cooler
- Asus z170 Pro Gaming Aura mobo
- i7-6700k processor
- 750 watt evga gold g2 supernova psu
- 16gb evga 3200 ram
- Asus strix 1070 graphics card
- Asus wifi/Bluetooth card
- Samsung 960 Pro 512gb ssd as boot drive
- WD Black 2tb hdd as data drive
- LG BD disc reader/writer

I haven't done any overclocking except turning XMP on for the ram.

Chances are your board is feeding too much voltage with the auto setting. You'd probably be best going into the bios and reducing this by perhaps setting an offset of -.1v? I think is somewhere to start. I'm not sure of the average stock voltage for the 6700k though.

The 212 Evo should be perfectly adequate for the 6700k at stock clocks.

I'd also consider re-mounting the cooler just in case.

VID way to high for stock! No mystery here. Configure your CPU vcore to manual instead of auto and try say, 1.25 to start with. Then boot into Windows and check the VID again in HWMonitor or CPU-z. Make adjustments as necessary to find the lowest manual voltage that will give you stability.

Thanks for the replies. I tried running prime95's blend test again after setting the vCore to offset by -.1 and it's now holding around 1.24 vCore at idle and temps have dropped to 81 degrees maximum.

VID way to high for stock! No mystery here.

Click to expand...

Is there anything I can do to change the VID or is that indicative of a bad processor?

Also, I'm still hitting max vCore during the blend test of 2.752. I've read that the max vCore should be much lower than that. Is this related to the VID?

This is HWMonitor after running prime95 blend just now:


Thanks again for your help

VID is what the CPU thinks it wants. For example, my main 6700k system right now, at fixed all cores 4.2 GHz is showing up to 1.40, and my 2nd system at same settings is showing VID up to 1.33. I'd suggest ignoring it.

What's more important is what the mobo is supplying. Many have an annoying habit of enabling an "enhanced turbo" mode which is essentially an overclock, and they usually give it more voltage than necessary while at it. Also be wary about software reported voltages, since they may or may not be reading what you think it is. I found setting 1.25V to be adequate at 4.2 all cores on my samples, but they can go Prime95 unstable below that.

First thing I would try is updating the BIOS. Could be bugged.

That is very weird that hw monitor is saying your max vcore is 2.9v Regardless even if your cpu is getting a constant 1.45v it shouldn't go to the mid 90's. Disable the svid setting in bios and see if it still reads that way. I know hw monitor says it's getting a max of over 2v but I don't think that is correct. Maybe your thermal paste needs redone. Try download Aida64 and see if the monitoring tool in that program reads the same while you use it to stress test. It is usually pretty accurate. I've seen a few cases where hw monitor gives false readings. But not very often. I have never had an issue with it but some people do. Asus boards are notorious for high voltages as well so I always advise if you use a z170 Asus board just go in and set it to manual. Might as well do a quick overclock to 4.5 Ghz as well because Skylake is one of the most consistent overclocking chips there are. You are pretty much guaranteed to get a 4.5 Ghz overclock rock solid stable at under 1.35v. And 100 percent guaranteed to get 4.5 under 1.4v Aida 64 also will give you lower temps stress testing than p95 especially if you're using a version later than 26.6

Current BIOS for this MoBo should be 3016. It was updated on Dec 29 2016

This system should be running stock with no messing about. Something is amiss.

I believe recommending an overclock before figuring out the problem is bad advice. IMO.

May be worth just resetting the bios to optimized defaults and see what you get. The vcore number of 2.994 is not right at all, no cpus call for that much voltage. As soulcatcher noted, are you running on version 3016?

Also, is that the latest version of hwmonitor?


@Rainwater, while overclocking advice is fine, this is not the time for it. It's important that the OP gets their computer stable and to the point it needs to be at stock before even considering messing with overclocking.

Ok guys! So I downloaded AIDA64 and ran it side-by-side with HWMonitor while running Prime95 small FFT test and the results were strange, but good I guess. I noticed that HWMonitor showed exactly TWICE the Vcore voltage reported in AIDA64, as shown in the picture below:


So apparently my vCore is only hitting a max of around 1.312 and not 2.624.
I'm currently using HWMonitor version 1.30.0 Driver version 140, which I believe is the latest version.

Also, I used Asus EZ Flash BIOS updater and it said my BIOS was up-to-date at version 3002... so thank you for pointing out that this is incorrect. I will update my BIOS and then re-run the tests and let you guys know what happens.
One thing that still bothers me is that my temps are still hitting mid-to-high 90s during Prime95 testing even though I have all fans set to max. I will definitely be reapplying thermal paste because I plan on replacing this cpu and mobo anyway since I'm still within the return period.

Disable the svid setting in bios and see if it still reads that way.

Click to expand...

I will try disabling this option in BIOS if the update doesn't help.

Ok I updated the BIOS to version 3016 and I didn't have any changes. So I went into BIOS and reset optimized defaults, which set the vCore to 1.35v. I then set vCore to manual at 1.25v and AIDA64 is showing that it is indeed staying right around 1.25v and goes up to between 1.28v and 1.296v. So I THINK that the voltage problem is solved, however my prime95 temps are still at around 90-91 degrees.

Please let me know if I might be missing some other settings or if I just need to go ahead and start disassembly to return mobo and cpu.
Another question I have that I know there is a lot of debate about is what method should I use to apply the thermal paste? I used the spread method for this build, but I am thinking about using the dot method with a thick blob for re-assembly.

I generally stick with a small x in the middle or a blob the size of a small pea in the middle of the cpu and let the heatsink do the work of spreading it out.


My only other guess is that somehow you are getting very poor contact with your heatsink, I highly doubt that your cpu and or mobo are malfunctioning (possible, but I doubt it)

I generally stick with a small x in the middle or a blob the size of a small pea in the middle of the cpu and let the heatsink do the work of spreading it out.

My only other guess is that somehow you are getting very poor contact with your heatsink, I highly doubt that your cpu and or mobo are malfunctioning (possible, but I doubt it)

Click to expand...

Thanks, but do you think that the high temps and voltages could have reduced the lifespan of my cpu and/or mobo? I am thinking it might be worth the extra work to remove and replace them
just in case.

VF-25_Messiah said:

Thanks, but do you think that the high temps and voltages could have reduced the lifespan of my cpu and/or mobo? I am thinking it might be worth the extra work to remove and replace them
just in case.

Click to expand...

No. Not to worry. TJMax is around 100c. You would have to run high temps for a long long time to shorten the life of the CPU. I certainly would redo the TIM and the mount. And tell us about your overall case ventilation.

The temperatures should be fine with stock Bios settings, I would check the Heatsink fan mount and make sure the fan is operating correctly.

Ya I know you're right. A 300 Mhz overclock is severely dangerous. None the less if it came off incorrectly excuse me. I wasn't suggesting that be the first utmost important thing he does. I gave a few ideas to try so he can make sure it's a false reading. I was simply letting him know what he can expect from Skylake. A lot of new builders are intimidated by overclocking. I always encourage getting a little extra performance out of your chip. I didn't say shoot for the moon, shoot for a very mild overclock and guaranteed at that speed he could do it with a safe voltage. Now temps, on air some versions of p95 are going to give you high temps. Glad Aida helped put your mind at ease with voltage. Cpu-z tho it is also a cpuid software might give you an accurate real time reading too I hope. If you use Aida to stress test instead of p95 since you are on air it may give you lower temps. Me personally yes, if you run Aida as a stress test and hit 90c would re apply your thermal paste. But that is me.

It was clearly obvious there were false readings, and again I would never try and get anyone to damage there system lets not start the new year off shedding too many tears now.

My heatsink fan is making a noticeable quick grinding noise when ramping up above and down below a certain rpm so I'm going to remove the heatsink and return it along with the processor just in case. Hopefully this will fix my problem.
By the way, I had to replace my first mobo (which was the same as the one I have now) after installing all my drivers due to a bad usb 3.0 port that I didn't check until after I had installed Windows and all drivers. Should I have uninstalled/reinstalled the drivers after replacing the mobo?

And tell us about your overall case ventilation.

Click to expand...

My case config is in my signature. The only thing I noticed about the chassis fans is that the two front intake fans only reach up to around 850-900 rpm with the define r5's front panel closed; while with the door open they hit their max rpm of around 1000. I ran prime95 with and without the front panel open and didn't notice any significant changes in temps after running prime95 for 5 minutes. My temps for the blend test are now reaching 84 degrees after 5 minutes rather than the initial 94 degrees within a few seconds. The small FFT test is still hitting low-to-mid 90's though, hence the above-noted plans to replace heatsink and processor.

Thanks again guys for enlightening me as to what is going on with my cpu and ideas on how to fix it. I originally didn't have any intentions of overclocking right now, but after having to get into some of the specifics of it I'm thinking of trying a mild overclock now (as long as I can get the damn thing stable at stock speeds!)

If it is version 28.xx 84 on blend is about you can expect. Try running Aida stress test instead and see where your temps are. Your cpu should be totally fine. If you're having trouble with your cooler then who knows maybe that is why it's getting so hot. I mean if where ever you got the cpu is willing to just give you another then i guess it won't hurt but it will throttle and shut down before any real damage will occur due to heat. With an air cooler you might be better off making sure it is an older version of p95 like 26.6 if you're going to stress with it. And usually the fft test is to test your thermals anyway. Blend is better for stability.