I7 6700k running WAY to hot and I have no idea why!

[SonicMango]

So I have a 6700k and CoolerMaster Hyper 612 v2 for my heat sink. I'm idling in the mid 20s and general use gets to around high 20s-low 30s. I Went into RealBench and went to do a benchmark on it because I wanted to start overclocking it and wanted to see how it handled things at base. So it near instantly shot up to 92c on the encoding section. I immediately shut off the test and it went down to the norm again.

I found out that I used to much thermal paste my first time putting it on, so I got it all off and reapplied. No change in Idle and during the same test it went up to 88c but got there slightly slower. So I thought 'ok maybe still too much paste', I had seen you should basically use as little as possible and 88 still sounded very high to me. So I just finished taking it all off and reapplying even less so that now its a very very thin layer on my processor and my idle temp went up about 2 degrees and it's jumping around between 26 and 40 seemingly erratically.

I can't find much else and I had gotten very good and helpful answers here before so I ask you, what am I doing wrong? I really have no more ideas on what this could be and why it's running so high.

 

[jollypirate]

How did you apply the paste? did you spread it out on the cpu?

 

[SonicMango]

How did you apply the paste? did you spread it out on the cpu?

I placed a couple of tiny blobs around the center then spread it out with end of the applicator. Not corner to corner though because the sink would squish it down and spread it out.

 

[Sergeant_Sneaky]

How did you apply the paste? did you spread it out on the cpu?

I placed a couple of tiny blobs around the center then spread it out with end of the applicator. Not corner to corner though because the sink would squish it down and spread it out.

Multiple blobs may cause "air bubbles" try a single pea sized dot and don't spread it, and see if that improves things, you can find a video guide on youtube to get an idea of the size.

 

[SonicMango]

Multiple blobs may cause "air bubbles" try a single pea sized dot and don't spread it, and see if that improves things, you can find a video guide on youtube to get an idea of the size.

I did watch a video on it and placed a spot right in the center and the temperatures got slightly worse. Right off boot it was nearly 50 and idles between 28-32 now and jumps between those. Like I have Firefox open making this reply and having HWMonitor and Speccy open and it's jumping between 45 and 55.

 

[einzele]

how much do you oc your chip? do you overvoltage too much? iirc 1.4v is the most u should set at. the funny thing is your idle temp is at freezing 20s.. it could be a bad chip u have there.
as reference my 4790k @4.4GHz on nzxt kraken x61 in air-starved H440 case hits 30+ on idle, and 60+ on 100% load prime95.

 

[SonicMango]

how much do you oc your chip? do you overvoltage too much? iirc 1.4v is the most u should set at. the funny thing is your idle temp is at freezing 20s.. it could be a bad chip u have there.
as reference my 4790k @4.4GHz on nzxt kraken x61 in air-starved H440 case hits 30+ on idle, and 60+ on 100% load prime95.

I haven't OCed at all. Its at base everything. That's why I'm so confused because in the mid 20s is a really cold then it would shoot up to the 90's under load. I was able to make an exchange for a new one with Amazon and it should be here tomorrow.

Just a thought, since Skylake has the voltage regulator on the board and not in the chip could it be a faulty board by chance? I wouldn't think it would be my heatsink There are some grooves between pipes but they all get filled in by paste and I dont see any scratches along the copper.

 

[einzele]

how much do you oc your chip? do you overvoltage too much? iirc 1.4v is the most u should set at. the funny thing is your idle temp is at freezing 20s.. it could be a bad chip u have there.
as reference my 4790k @4.4GHz on nzxt kraken x61 in air-starved H440 case hits 30+ on idle, and 60+ on 100% load prime95.

I haven't OCed at all. Its at base everything. That's why I'm so confused because in the mid 20s is a really cold then it would shoot up to the 90's under load. I was able to make an exchange for a new one with Amazon and it should be here tomorrow.

Just a thought, since Skylake has the voltage regulator on the board and not in the chip could it be a faulty board by chance? I wouldn't think it would be my heatsink There are some grooves between pipes but they all get filled in by paste and I dont see any scratches along the copper.

the board could set the voltage higher than it should, but you can check the voltage with monitoring software such as CPU-Z. 1.2V is the normal value. if u see the voltage higher than that, u can change it manually under bios.
it could be a faulty chip, lets hope you have a good luck in silicon lottery for the new chip 😛
also, i recommend just put a pea-sized paste right on the middle, and let the heatsink pressure spread it evenly, it works better that way in most case since it minimize the chance of air bubble trapped in between. remember the "less is more" rule also 😀

 

[SonicMango]

how much do you oc your chip? do you overvoltage too much? iirc 1.4v is the most u should set at. the funny thing is your idle temp is at freezing 20s.. it could be a bad chip u have there.
as reference my 4790k @4.4GHz on nzxt kraken x61 in air-starved H440 case hits 30+ on idle, and 60+ on 100% load prime95.

I haven't OCed at all. Its at base everything. That's why I'm so confused because in the mid 20s is a really cold then it would shoot up to the 90's under load. I was able to make an exchange for a new one with Amazon and it should be here tomorrow.

Just a thought, since Skylake has the voltage regulator on the board and not in the chip could it be a faulty board by chance? I wouldn't think it would be my heatsink There are some grooves between pipes but they all get filled in by paste and I dont see any scratches along the copper.

the board could set the voltage higher than it should, but you can check the voltage with monitoring software such as CPU-Z. 1.2V is the normal value. if u see the voltage higher than that, u can change it manually under bios.
it could be a faulty chip, lets hope you have a good luck in silicon lottery for the new chip 😛
also, i recommend just put a pea-sized paste right on the middle, and let the heatsink pressure spread it evenly, it works better that way in most case since it minimize the chance of air bubble trapped in between. remember the "less is more" rule also 😀

Yeah I hope I win big! Sure did with my graphics card! Things a champ!

 

[Lars01]

hello, i also got a problem. i have a i7 6700k with a noctua nh-d15 and got 78 degrees on prime95. Idle i get 23-25 degrees in a 23 degrees room? Faulty chip?

 

[einzele]

hello, i also got a problem. i have a i7 6700k with a noctua nh-d15 and got 78 degrees on prime95. Idle i get 23-25 degrees in a 23 degrees room? Faulty chip?

a bit higher than i would be comfortable with, but nowhere dangerous. dont worry about it too much, prime95 gives your cpu unrealistic work, producing unrealistic temperature.
your cooler is great though, it should be on par with top performer coolers on the market.
check your thermal paste, and how did u apply them? i recommend put pea sized blob in the middle and let your heatsink spread it evenly. also check your cpu voltage, 1.2 should be the default number.

 

[Lars01]

Ok, i am going to re apply the thermal paste when i come home. I actually yhink i put to little thermal paste. I am also going to set down the volt if it is too high. If these thing's dont work, i am going to send it in and get a new one (if i can). You see, i want to overclock later on, so i want a cpu with low temps. That's why i bought a nh-d15.

 

[leonardo2209]

I also have Probems with the temps, I have an alphacoool Eisberg 240 with 2 Corsair SP120 on it.
In Idle it is about 20-30°C but sometimes there are peaks over 65°C+
sometimes just by opening an folder or firefox or something like that
Running prime95 it goes instantly to 88°C
The Pump is contolled by my Asus Z170-E Board.
My 2500K before was on prime95 never above 55°C
Sorry for my bad Englisch 😉

 

[Loack]

Hello, I have exactly the same problem: my CPU has a fine temperature when idle, but it will spike under a 2-12% load up to 55 degrees Celsius and it will easily achieve 78 Celsius degrees under RealBench (without any kind of overclocking). Changing BIOS settings and voltage doesn't change things much. I must add that pushing it slightly (up to 4.3 Ghz) doesn't change much either : the CPU just gains a few degrees, which is expected... I tried to reapply thermic paste and even to change it for a better one (Noctua)... My cooler is a All in one Cooler Master Seidon 120V Ver. 2 (not the best one, but previously it was very comfortable with my i7-2700k overclocked to 4.6 Ghz). Given what I can read here and there, I now think that there is definitely something wrong with *a lot* of these CPUs (the i7-6700K). Intel seems to stay silent about these issues and I haven't found (yet ?) many articles from the Press regarding it (there are a few however, that point to the lack of performances and excessive heat/voltage...)

 

[Ecky]

There's nothing wrong with these CPUs. Modern Intel CPUs can safely run up to around 100c (varies slightly depending on the model). You can't compare your expectations of what older processors would run at.

There are at least 20 posts per day in this forum along the lines of, "Help, my Intel CPU is within 20-30c of its max operating temperature when I stress testit, what do I do?", to which the answer is generally, "Nothing, that's normal".

 

[Loack]

Ecky, can you please explain how you find this: "Modern Intel CPUs can safely run up to around 100c (varies slightly depending on the model)."? Is it by experience or is there an official statement confirming this? And how do you precisely define the "modern CPUs": is this the Skylake generation? Or Broadwell ? Or Haswell ? I am very interested in your advice, as I thought that 70-75°C was an upper limit for proper operation. Don't you think that a CPU operating at 85-95°C will be damaged in the long term?

 

[Ecky]

This was written regarding older Core2 processors, but the basic concepts (if not the exact numbers) should still apply:

http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf

Functionality, electrical specifications and reliability commitments are guaranteed at maximum Tj as measured by the DTS.

Tjmax for Skylake is 100c, which can be viewed by utilities such as CoreTemp or RealTemp. They'll have "core temperature" which is basically a reading from the digital thermal sensor (DTS), "distance to Tjmax" which is how far you are away from maximum temperature before throttling, and "thermal junction maximum", which describes the maximum temperature the CPU will operate at before throttling. Emergency shutoff is generally closer to 130c.

 

[Loack]

This was written regarding older Core2 processors, but the basic concepts (if not the exact numbers) should still apply:

http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf

Functionality, electrical specifications and reliability commitments are guaranteed at maximum Tj as measured by the DTS.

Tjmax for Skylake is 100c, which can be viewed by utilities such as CoreTemp or RealTemp. They'll have "core temperature" which is basically a reading from the digital thermal sensor (DTS), "distance to Tjmax" which is how far you are away from maximum temperature before throttling, and "thermal junction maximum", which describes the maximum temperature the CPU will operate at before throttling. Emergency shutoff is generally closer to 130c.

Hi, I agree with your point of view, but it is clearly stated that this is a maximum (tjMax). For proper, sustained, operation with proper cooling, I think that the acceptable temperatures are lower: we should be looking for TCase, no? For instance, I found on this very website (http://www.tomshardware.co.uk/forum/id-1800828/intel-temperature-guide.html):

Section 6 - Throttle Temperature

Also called "Tj Max" (Tjunction Max), this is the Thermal Specification that defines the Core temperature at which the processor will Throttle (reduce clock speed) to protect against thermal damage. Although Intel processors are capable of operating above 90C, we also know that excessive heat kills electronics.

Sustained Core temperature greater than 80C is too hot for ultimate stability, performance and longevity.

And also:

At Default / Auto BIOS settings (stock clock and Vcore) with 100% workload, Core temperature is 5C higher than the Tcase specification - http://ark.intel.com/ This means that whatever the Tcase specification is for your processor, add 5C to get the corresponding value for Core temperature.

For Intel I7-6700k, TCase is 64°C, so proper operation should be at 68°C. I am tolerant: I would accept 70°C.

Also, I think that this is a bit excessive to sell CPUs as unclocked if they can't be overclocked without having a custom WC loop (due to excessive heat and, consequently, noise)? Don't you think so? I would rather think that Skylake high thermal profile is not a feature by design, but a flaw of the architecture...

Happy to read your advice.

 

[Ecky]

What should be 68c, though? Tcase is not core temperature, it's temperature at the top of the heatspreader. There is no way to read what this temperature actually is, as there is no Tcase sensor, and I was unable to find any place on Intel's site that actually says that Tcase is 5c lower than core temperature. Tcase was *probably* around 5c lower than core temperature back in the Core2 days when heatspreaders were soldered onto the top of cores, but it's a fact that there's a larger gradient between core temperature and heatspreader temperature since Ivy Bridge, when Intel stopped using solder and instead began using thermal paste.

Mobile CPUs use Tjunction for maximum temperature ratings, because they don't have heatspreaders. There's no reason a desktop CPU would be any different.

I agree that when overclocking/overvolting, it's wise to have some thermal headroom. A combination of temperature and voltage are what cause degradation of CPUs, and if you're raising voltage, you probably should compensate by lowering temperature. But then, you're running the CPU out of spec and a lot of things change.