Thermal sensor issue i7-7700k?

[CK7]

I have a brand new build; EVERYTHING NEW. i7-7700k is running at stock speeds. I have the RAM set to XMP for DDR4-2666. Motherboard is Asus Maximus IX Hero Z270.

I have found that the i7-7700k reports a momentary (a second or less) temperature spike +25 > 35 degrees Celsius anytime a program is opened, a webpage is opened, a background app runs etc. The temperature blip cascades through the cores in random order; not the same every time. This causes my heatsink fan to constantly cycle up and down. Temperatures otherwise report as steady, normal increases. Peak temperature under Prime95 blend test is 71 degrees Celsius.

Attempted solutions:

I have re-installed my heatsink and thermal paste with no change.

I have tried to manually set my fan speed in the bios. The only setting that avoids this issue is setting the temperature / fan at a constant (and loud) 80-100%. I've tried PWM and DC mode.

I have found a few user reports elsewhere on the web, all reasoning that it's just the way it is. I don't accept that. Opening a folder or browser should not spike temps +30 degrees. Not only is the fan cycling annoying, it puts undue stress on my fan; possibly shortening its lifespan.

What's the answer, if any? RMA?

[TGrab1]

RitchieDrama wrote:

Disagree all you want, anyone with a Z270 Gaming Pro Carbon has no issues.

First off you and a friend are a very small sample size in the scheme of things.

Second I am sure I saw dumps of MSI boards at the beginning of this thread when we were submitting information to Intel.

So your result are not indicative of everyone.

[idata]

TGrable wrote:

I have to disagree that it is board related.

Some boards were having issues with auto voltage being too high, but bios updates fixed most of that.

If you go with a solid name manufacturer though then the board should all be very similar minus certain feature sets etc.

I personally use an ASUS board and have had no issues with it. So I will say though that if you are running a board on auto.. make sure you get the newest bios updates for it.

Disagree.

My Asus Maximus Gene VIII motherboard has the very latest BIOS update, still had problems with the default BIOS settings of "auto". I always update the BIOS to the latest releases.

[idata]

I'd say 3 7700K's and 3 Z270 MSI Carbons would be a very clear indication that I know how to set the voltages up and they're fine.

I've also helped people on this forum, reddit and overclock.net - they have all achieved the same result as me due to it.

All I have seen from you in the last few pages is you trying to talk about something while being out your depth with very little understanding.

[idata]

I'm running Prime version 29.1 Build 4.

From my rather limited experience I would say that Prime95, small FFT, is a very good stress test. But as I mentioned in my post I was also running AIDA54 stress test and MemTest86 multiple copies, AND playing a video, plus normal work ... so I think the CPU and memory were being pretty well exercised. As the temperatures were at under 65C after several hours testing I think it's fair to say that a manual setting of 4.5GHz at 1.155v is very OK on this particular CPU, on an ASUS Prime Z270-A.

Out of interest I'll try clocking up towards 5.0GHz to see how the temps look, although I've no real interest in running at that speed.

Cheers

Robert

[idata]

ra5040 wrote:

I'm running Prime version 29.1 Build 4.

From my rather limited experience I would say that Prime95, small FFT, is a very good stress test. But as I mentioned in my post I was also running AIDA54 stress test and MemTest86 multiple copies, AND playing a video, plus normal work ... so I think the CPU and memory were being pretty well exercised. As the temperatures were at under 65C after several hours testing I think it's fair to say that a manual setting of 4.5GHz at 1.155v is very OK on this particular CPU, on an ASUS Prime Z270-A.

Out of interest I'll try clocking up towards 5.0GHz to see how the temps look, although I've no real interest in running at that speed.

Cheers

Robert

Run Prime95 Small FFT's test on its own - nothing else except the temperature monitor. Close everything - even those applications in the system tray (bottom right area, often hidden behind the chevron). I found Prime95 will do full stress testing when no other applications are running, otherwise the other apps will interrupt Prime95 and it won't get 100% of the CPU attention. You can still monitor the temps.

[idata]

Yes, good ... that pushes it to 71C max (at 4.5GHz and 1.155v) and solid on small FFT.

I also tried pushing the clock up to 4.9GHz. At that I need 1.34v to get a stable Prime95, but the temperatures are up in the low 90s, which is not comfortable for me. I also don't see the point as all I would be getting is an additional 9%, which I very much doubt I would notice.

[idata]

Okay, well you shouldn't be running that version of Prime95 due to AVX.

ftp://mersenne.org/gimps/p64v265.zip

Run this one only.

[idata]

Why would you not want to test AVX, since the processor supports AVX? It's true that the applications I use don't currently use AVX instructions, to the best of my knowledge, but future releases may, so I would be better making sure the processor is stable with AVX.

[idata]

For the same reason that your bios has (or should have) an AVX off-set. Why would you stress test your PC with AVX when 99.9% of the stuff you use, will not use it.

Please be realistic.

[ASušt]

The thread growns impatience...

My 2 cents about voltage:

I use adaptive mode for vCore.

So in 90% of time when i'm surfing the web my cpu is at 0.8v and ~1.6-2.5Ghz....and it spikes 60c.

About cooling:

It's cool that some of us have custom water parks to hold everything under 30c.

But there are people with....(drumroll).....stock intel coolers. Why not? 7700k are provided with those coolers.

So if e.g I have my Idle state at 30c, they have it ~40 maybe idk. Mine spikes to 60c, and theirs? 80c?

I'm sure, it's not a problem of a weak cooler.

In my previous PC there was i7 3770k @ stock.

I played for 2 hours, everything warmed up and when i closed the game, everything was starting to cool down. Slowly.

I could clearly see that my rig is warming or cooling. Not that it's totaly cool, second after - overheats..no it's not.

And for RitchieDrama and TGrable

Guys, it looks like you are on Intel's side.

If everything suits you and you are ok with this situation, please, let us get our answer from Intel. Use your MSI boards and deliding tools all day long and be happy.

TGrable, thank you for all the info you provide, but that still is your theory. Pity that you won't go to my place to delid my cpu, and probably Intel won't do that to.

I hope we won't start a world war here, while we wait intel to reply to us.

[TGrab1]

No war here.

I'm not on intel's side either. I am just trying to play devils advocate while also looking for anything else that could be wrong with a persons setup to try and troubleshoot that first.

If we bring Intel a bunch of data that just shows it is user error that will not further are case and I'm not even saying most of it is that, but if we can try to troubleshoot it ourselves here first all while in the view of intel and they can see..

They tried this, that, this, and also a little more of that and the problem still persisted to a degree even after all of those steps... then maybe we will be on the right path to getting solid answers as we have tried all the other alternatives.

Me trying to ask questions and offer advice or point things out isn't done so to make intel look better, but instead to make sure that we all have done everything we can outside of replacing the cpu.

Also to whoever said I was out of my depths, I am sorry you feel that way. I just don't always have enough time to type out as much detail as most would like or maybe even as I should.

Lastly, Yes I always test with p95 with avx 2.0. The reason behind this is I want to see worst possible situation. P95 is harder on the chip than almost anything else you can throw at it.. So if I can pass my own p95 stress test with the temps under control and for the duration I am comfortable with so be it. Just means I am confident in my overclock and the worst case temps I would see. There are other alternatives, but I have found they either do not push the temp envelope as hard (which I like to test) or the test takes much longer to fail to give you a good idea of a stable overclock. Try running p95 with an unstable clock and 99% of the time a core or two will fail within the first 30 seconds. That isn't the case with Aida and some of the others.

Then again it all comes down to personal preference. Once again I am here only to try and help or offer any knowledge I have on these areas. You can attack me if you want, but I am not here for intel or anyone. I am here because I am bored and looking to help in my off time.

[ASušt]

Ok, so no offence.

Good thing we understood eachother.

So, if you are offering advices, what do you think about 4,8ghz at 1.25v in adaptive mode? Works pretty stable, but maybe you have something to add?

[idata]

If you really want to push your CPU then I think the way to do it would be:

- find out the max settings that your CPU is stable at (say 1.34v at 4.8GHz) WITH AVX stress testing.

- find out the max settings that your CPU is stable at at that voltage (say 5.0GHz) WITHOUT AVX stress testing.

- run your CPU at 5.0GHz and 1.34v with an AVX offset of -2.

That way you should have a stable system capable of handling anything that is thrown at it.

In my case, as I'm not interested in running much faster than 4.5GHz I use a slightly different approach. I find out the lowest voltage that the CPU is stable at with AVX at 4.5GHz (in my case it's 1.155v). I then find the highest frequency that it is still stable at without AVX at that voltage (in my case 4.6GHz). I then set the CPU to 4.6GHz and 1.155v with an AVX offset of -1.

And btw, the latest version of Prime95 has the option of turning AVX on or off through the local.txt settings.

Cheers,

Robert

[idata]

Here is a response from Intel customer support to the issue with my CPU:

"Thank you for your response. In order to evaluate the issue you are experiencing we need to know the information asked to you on my previous email. Please note that, within the technical department, we haven't got any reported overheating issue related to this CPU and, therefore, we need to know every specific case to determine the best way to assist you further. In order to do so, please answer the following:"

So it seems that at the very least we should put a support call in to Intel Customer Support.

Robert

[AOhan]

Processor Model: Intel i7 7700k 4.2ghz

Motherboard Brand: Asus

Motherboard Model: STRIX Z270F Gaming

BIOS Version: 0505

RAM Part Number: CMK16GX4M2B3000C15

OS installed: Microsoft Windows 10 pro 64 bit

Video Card: GTX-1070-GAMING-X-8G-MSI

I have Noctua D15 Cpu Cooler.

my cpu lowest temp is 32-33c and its spikes to 60-70c in idle even without doing nothing.

at load it reach up to 88c.

[idata]

Iam getting so pissed of this Intel bullshitting..I feel robbed by Intel..makes me so freaking pissed of this company..I put 3500€ for this computer and only what i got is freaking temp problems and other problems..and only solution for this is to delid my cpu and lose my warranty..

all I can say here is fuck off Intel and stop robbing your customers or soon you dont have them anymore!!

And this answer..what a bullshit is this..lol..its normal..its not fucking normal !!! I think I am done here...

[idata]

A lot of this is to do with motherboards.

I have my CPU at 1.328v, 5ghz, and my temperatures don't go over 73 when gaming, which is absolutely normal and fine. I get the odd spike when on desktop (rare) - I idle depending on time of day between 24-32, if it does spike for a second, never past 55.

I can run Cinebench R15 CPU test max temp usually hits between 69-72, I can run Prime95 max temps usually about 78 -- I don't run it for long though as its a stupid way to test.

I am not de-lidded either. H100i V2 with pre-applied paste.

I've setup my friends 7700K, same bios settings as me, 5ghz etc, same temps as me.

[idata]

Hi

I also had this issue on my i7-7700K all running at stock values no overclocking but the temp never went over 80 °C so I did not care much about it. I cannot however provide diagnostic data because my CPU died last Friday while I was away from my PC. I tried it in 3 different Asus motherboards and in everyone the boot process stops after 2 seconds with code 40, which is according Asus manual ACPI checkpoint - System is waking up from the S4 sleep state. The CPU was in use for 2 months.

[nx1]

guys.. without relating to the OC\Cooling issue - on stock settings the i7 7700K is spiking.

if you open up a browser, clicking on something and general normal usage on Windows - its not

supposed to spike for 1-3 seconds and cause the temperature to rise for about 20-25C instantly.

the big question : is it fixable with BIOS update ?

Intel isn't answering ATM, and im not seeing any news about it from motherboard manufactures.

if im looking at the LONG RUN - i don't care if these spikes\temps are SAFE. its not supposed to happen,

and most definitely - im not supposed to hear my Noctua fan go nuts for a few seconds when im replying to this topic for example.

it happens ALL THE TIME. im paying for a decent cooling for its qualities and - silent work on IDLE is one of em.

im giving it 2-3 weeks at best, if there won't be any legit answer\update on this issue, Ryzen is fine be me.

il even accept the loss of FPS ( in gaming ) and will never look back on Intel ( not ssd, not cpus, not even motherboards with LAN chips from Intel ).

[ASušt]

cohr

Count me in.

And touching the voltage again. In adaptive mode/balanced powerplan CPU drops it's voltage and clockspeed at idle.

So when my cpu spikes in idle, in that moment it is usually at less than 2Ghz and 1v.

So I think, I can put the end of thinking, that my CPU is overvolted or too much OC'ed.

Starting to hope, that my CPU will die like Marwyn's one did, so I could return it...

[LRibe2]

Hi once again guys!

I think we are discussing nothing that has an answer....

The thing is, this spikes stuff is a subject that has been in a lot of threads over the years now. And I've noticed that it happens in previous generations, and the only usual similarity is that all the CPUs that show this fast temp spikes are CPUs running above 4Ghz.

The real question is, is this a problem or a feature? Intel dosent seem to really care, and there arent any reports of malfunctions related to this.

The higher temperatures is another situation. I've questioned Intel in another thread about temperatures reaching 80C + temps while running avx instructions (Like prime95) and the answer was that, it' ok, the cpu is supposed to run hotter and it' supposed to reach 80C or higher while running this tests and go back to normal the moment you stop the test. Intel says it's safe and people are really making to much "noise" about it while it's not a problem.

I've OCed my CPU manually to 4.7Ghz and an AVX negative offset of 2 and Vcore at 1.23V (Reaching max 1.248V), so that it runs at stock on avx instructions. It's stable but I have my limitations using a normal cooling system in an mATX case (Evo 212 and bitfenix prodigy). So I can't push it above this in a safe way.

This CPU is a beast and besides Marwyn, I havent heard about any broken or fried 7700k.. And I havent also heard about any 6700k that's malfunctioning or a 4790K and so on... and this CPUs also have "the spikes problem"

The CPU has a 3 year warranty.. I think I'll wait to see what happens but I bet that 5 or 6 years from now, It will still be a beast and working just fine..