How best proceed with overheating i7-4790K?

[REnso1]

I have an i7-4790K in a GA-Z97MX Gaming 5 mobo latest F4 BIOS.

I used a Noctua NH-L12 cooler rated at 95W for cooling the processor rated at 88W in a Lian Li PC V354 with 4 fans, 2 in 2 out, case closed and case open, ambient is 27-30°C.

Temperatures in BIOS and memtest86+ were high so I decided to try stress testing and in Prime95 small FFT cores 1&2 overheated to 100°C using Core Temp.

I tried reseating the heatsink and renewing the NT-H1 TIM and opening the case but it made no difference. I have a photo of the contact pattern here.

When I tested using the OCCT benchmark I was unable to complete a test due to the processor overheating so I underclocked the processor to 3.6 GHz, disabled turbo and manually set vCore to 1.1v.

With an underclocked processor I was able to get a heating and cooling curve using the OCCT auto capture, to enable me to study the problem.

Even when underclocked the processor was reaching high temperatures, rapid fluctuations in temperature with work load suggest a bottleneck in the thermal pathway. When I tested with the intel retail cooler which came with the CPU the cooling was much less effective than the NH-L12 (even when underclocked taking just over a minute of OCCT to reach the 85°C cut off point see below) indicating the NH-L12 was doing a good job of removing heat, which meant the processor was making the heat or the source of the bottleneck.

I have discussed it http://forums.hexus.net/cpus/327593-4790k-overheating-nh-l12.html elsewhere. Advice was to contact Intel due to an absence of information relating to my retailer's testing procedures. I have asked about these but am still waiting for a reply.

So my question is how should I proceed from here? Does this qualify for an RMA? If so is it possible to negotiate this with Intel direct or do I have to go through my retailer?

I have done my best to make sure I am not doing anything wrong and I would be grateful for any pointers to any mistakes I may be making.

[oouim]

Hello, I have a problem with overheat , I can not play a game with the graphics to the lowest , I play two minutes , the processor gets hotter than 88c ++ ... I already made up a update bios, and everything that was in my power to repair my i7 , so I decide to opens a request for an exchange with intel ... but its been 15 days already I speak with an employee , and 4 days now I send my setup and my shipping information , and I have no answer !! this is normal ? how many months I have to wait? ?( I have just want their addresses to send it my i7 , why that take four days send a address? ) I pay over $ 2500 for my new computer I can not do anything with !! just because the i7 DEFECTIVE ... it's normal ?? how long its takes normally? why the service is so slow ? i just want my money or a other i7...

everyone spends a lot of time and money (new fan ect .. ) to solve this problem ! why we can not have a good service? I bought my computer there almost a month and I can just let him take the dust..

[WGual]

I am having the same issue. I have the 4790k in the GA-Z97X-Gaming 5. Occt fails immediately due to the heat of the processor. I am not an over clocker and have it all stock. I have updated the bios of the board to F5 the most recent revision as well. When doing any tests on the processor it goes to 95-100c instantly. I have used the Intel burn in test and the processor is way up there. Not sure what is happening. Seems that this is a know issue though.

[WGual]

My batch is L421B951as well. Hope it helps. I would like a solution to this issue. My processor is not even a month old.

[WGual]

I decided to replace the CPU. Will let you know if i have the same issue with the new processor.

[YNomi1]

OK you guys. I got interesting facts after 3 weeks of usage.

Without changing anything on my rig, I see small drops in temperatures every week. If you read my previous post I was having the same overheating issues with my 4790K as everyone else. Now I believe I have none. MAGIC …I don't think so.

First of all I run the IntelThermalTest v0.1 tool (BIG THANKS to karwos). Again let me remind you I don't use intel stock cooler. I use a Corsair H100i.

Here are my results:

this is about half way...

almost at the end...

and final results.

As you can see my temps never exceeded 52C. The same results i get with OCCT and AIDA64. The only program that gives me higher temps (about mid 80's) is the latest version of Prime95 which is known to cause this with the use of AVX commands. Let me stress that the test was done with 22c room temp and all the fans to 30% (the case was almost silent), so temps would be better if i used the 100% of the cooling capacity of my build. But i don't think this is the point here.

Anyway what has happened here is in my opinion what other users are referring to as heat spreader settling in. If the gap between heat spreader - die is bigger than it should, and this is causing the overheating, maybe then, if you apply adequate pressure with a good cooler, the heat spreader starts pushing the glue and "settles in" closing the gap between them. But that takes time. Because many of the guys here RMA the CPU almost immediately, they don't give the CPU the time to heal itself . Offcourse this doesn't change the fact that it's still a manufacturing error. Anything that comes out of a production line should work right away and without burn-in time. Also i am pretty sure now, that stock cooler is useless, and should be used, only if the CPU is used with turbo-boost disabled and ambient temps don't exceed 23C. They should have never given the 4790K bundled with a cooler, something which is misleading, and forces anyone buying this CPU spending extra money for aftermarket coolers.

Closing my thoughts let me say that during the test everything on my MOBO (GA-Z87X-UD5H bios F10c) was on auto.

[jband2]

do not know much about voltages but in the 3rd picture is that the voltages of + 3.3V , + 5V , + 12V , are lower when they should be higher than those voltages , will be well done the test?

[JVene1]

yianom wrote:

OK you guys. I got interesting facts after 3 weeks of usage.

If the gap between heat spreader - die is bigger than it should, and this is causing the overheating, maybe then, if you apply adequate pressure with a good cooler, the heat spreader starts pushing the glue and "settles in" closing the gap between them.

This is sound reasoning, presenting a hypothesis full of promise and ideas.

There are implications on many levels, including how enthusiasts (already decided on drastic action) might consider removing the IHS.

It would help to better understand what adhesive is used and it's properties. If this is a feature of the glue at reasonably low temperatures, like about 90 C, it may suggest one reason (among many) Intel would pause before issuing statements on the subject. If we imagine ourselves to be among the engineers, directors, VP's, legal counsel or any other members involved in making decisions on such a subject, this detail could provide a way to settle the problem without hinting at the need for a recall.

It's an uncomfortable proposal to say an owner should maintain a steady high temperature so as to heat the IHS to it's edges in order to heat the glue, but this is what your experience implies. I must assume that in the course of 3 weeks there were periods of operating in that temperature range, and other periods of rest. Perhaps that process could condense into a single run of some time which could range from an uncomfortable 24 hours to an anxiety inducing 96 hours or more. This suggests a "burn in" utility designed to push the processor just enough to obtain such temperatures, operating in the fashion of a thermostatically control heat source, to produce such results.

It also brings to mind one of the more absurd, yet occasionally functional ideas used to "repair" failed XBOX 360 devices....baking. The implication that heat might spread the glue under pressure also suggests that placing the CPU on a suitable protective surface, then placing it in a conventional oven controlled at a temperature below 100 C, would "treat" the glue to such a state that removing the IHS is several times easier than methods currently in vogue. It also hints there could be a type of clamp fashioned for this proposed process, which might squeeze the package during baking to settle the IHS onto the die.

There are many precedents for such ideas. An example is archived audio recordings on tape. Over time the glue which bonded the oxide coating to the acetate or Mylar backing would denature, becoming both sticky and gooey, allowing the recorded audio to lift off the backing thus destroying it. The solution is to "bake" the entire tape for a few hours at similar temperatures (below 100 C), which caused the glue to re-settle.

My point in that bit of history is that heat treatment of this type can be a double edged sword. On the one hand it can cause the glue to loosen, and then on the other, after it cools, the nature of the glue changes, possibly becoming even stronger.

This suggests that you may have been lucky. Other patterns of heating and cooling which fail to settle the IHS downward could actually make the problem worse, or making a solution to the problem by waiting less likely to succeed. A better understanding of the glue would help define that.

[idata]

To all who use IntelThermalTest and their tests being passed and under Prime95 it's failed

Can you please check, and re-run IntelThermalTest, and check in task manager (Ctrl+Shift+Esc) what CPU Load generates "IntelThermalTest.exe"?

It may be possible it didn't utilize full CPU time because i forgot to set each thread affinity mask to being executed by diffrent thread/core, so MAYBE Kernel OS for some reason executing some of threads sequentially on same core. Also i didn't increased default thread priority, this also may be cause.

So please check and let me know. It should oscilllate around 99-100. If it's 50-60, that's reason of PASSING the thermal test. I will relase 0.2 then.

Best Regards

[idata]

Becaue i've got confirmation that IntelThermalTest v0.1 was causing only 50-60% of CPU Load, here is new version:

=====================

= 28.10.2014 =

= v0.2 =

=====================

- Load threads have now set proper AffinityMask - CPU Wasn't fully loaded though

- Load threads have now set proper Priority - CPU wasn't fully loaded though

- SW can be less responsive now; after launching test just wait patiently 5 minutes

Download link is same, just re-download and post results : http://www.cyclone-networks.eu/IntelThermalTest.rar http://www.cyclone-networks.eu/IntelThermalTest.rar

Keep noted your PC after launching test may become less responsive (mouse problems, etc). Just ignore it - and patiently wait 5 minutes to get test finished.

[RKimb]

I'm puzzled by your test. I ran it and got temperatures up to 65C and was told the test was passed. However as I was pasting a screenshot Windows shut down with a blue screen saying that it was shutting down to protect my computer from damage. I then re-booted and ran 4 prime grid tasks and temperatures fairly quickly went up to 86C, at which point I stopped the tasks.

I'm not an expert, but it seems that your test and the prime grid tasks must be stressing the CPU in different ways. I wouldn't be very happy if Intel were to say that it's OK if my CPU passes some theoretical test, but overheats when confronted with real world computing.

[idata]

You ran test surely on v0.2? Can you post screenshot?

Because i have no any serious reply from Intel till now, and their support center is joke (asking same questions and dont read my mails porperly) im very seriously thinking about taking legal action. I have already gathered enough evidences.

Is there any from europe interested joining class action?

[RKimb]

Yes it was the version you posted a link to today.

As I explained, I had problems with the screenshot, but there was nothing remarkable - as I said it reported temperatures around 65C and said that the test was passed. I have a dual boot machine (I normally use Linux; I only use Windows when I have to) so I'd have to reboot to re-run the test.

[SBuck1]

I grabbed v0.2 this morning and tried it just to be sure. Passes fine @58C. However Prime95, at least the latest version, still runs waaaaay too hot. I'm gonna chalk that up to the weird instructions that v28.5 of P95 is using, and hope my cpu is fine.

[aanwa1]

Ok so these are the results of v0.2 alongside with task manager screen shot

and this is a screenshot when it finished

i also need to stress a point here after reading yianom reply, when i first built this PC 2 months ago i had a hard time installing my cooler Scythe Katana 4 and i needed to apply alot of pressure to install it.

when i ran prime 95 to see if the build is stable temperatures jump to 99 and 100 instantly my first thoughts were that i didnt install the cooler correctly but after i made sure it is (took it off and installed it many times) i had to think of another reason and i ended up updating my z97x gaming 5 bios to version F5 , after i did that temperatures stopped jumping to 99 when running prime95 instead they start from 70s and increase gradually till they reach 92 or 93.

with that in mind, i was sure that the bios wasnt optimized for this new cpu which is released after this mobo.another point that yianom mentioned is adequate pressure is good "the heat spreader starts pushing the glue and "settles in" closing the gap between them", when i looked at the back of my mobo it was little curved from the pressure the cooler applies, so is that maybe why my temperatures are better with karwos test than the rest here ?

[YNomi1]

When i first installed the stock heatsink that came with my CPU, i had Bios ver F9 (latest non beta bios for my GA-Z87X-UD5H) that was the first bios supporting the 4790K (and in general Haswell-E cpu's) for my mobo. When running Prime 95 or other test my temp from idle 35C, skyrocketed in just a second to 100C (and thermal throttling started kicking in).

Then i gave a try to F10c bios (which was stated as STABLE-BETA) and accompanied with a strange "Enhanced Intel K-sku CPU performance" tagline. When i installed the bios, i noticed that the CPU wasn't anymore skyrocketing to 100C in a second. It sure reached this temp with Prime95 (and only with prime95) but it did this more slowly and progressively like sometimes 1 minute or so.

So...

GA-Z87X-UD5H F9 (ALL STRESS TESTS) 35C ----> 1SEC ----> 100C ---> THERMAL THROTTLING

GA-Z87X-UD5H F10c (ALL STRESS TESTS - EXCEPT PRIME95) 35C ----> 1HOUR ----> 85C ---> NO THERMAL THROTTLING

GA-Z87X-UD5H F10c (PRIME95) 35C ----> 1-2 MIN ----> 100C ---> THERMAL THROTTLING

So surely there is a big connection in how the bios handles the new processor, but this is not the only culprit for high temps. I believe that with new bios versions from every mobo manufacturer we will see big improvements in temps. I am sure that in their secret underground labs, as we speak, they are doing tests trying to correct the problem as much as they can from their standpoint.

Also i must highlight again, that the above tests were done with stock intel cooler and 2-3 days after installing my new CPU. Being disapointed by the thermal performance of stock cooler i bought Η100i.

[RKimb]

"Because i have no any serious reply from Intel till now, and their support center is joke (asking same questions and dont read my mails porperly) im very seriously thinking about taking legal action. I have already gathered enough evidences."

I don't know much about the likelihood of success in class actions, but I'd have thought that some publicity might well do more good. The issue could be brought to the attention of computer journalists, Tom's Hardware, etc., etc.

[SElwa]

I have contacted Intel several times and I'd like to share the knowledge that I have acquired from them, which means It is official information not just guessing:

The first information which most of us knows: The CPU temperature (aka TCase or IHS temp) under stress test/load using the stock cooler is 74 degrees Celsius

The second information: It is expected to see a difference between the CPU temperature (aka TCase or IHS temp) and the core temperature of around 15-20 degrees Celsius.

The third information: it is expected to see a difference between cores of around 15-20 degrees Celsius.

The fourth information: If you tried everything to solve the problem, including updating the motherboard BIOS to the latest version and the problem continues, the processor should be replaced.

Notes: as we can see officially that processor under stress test/load using the stock cooler in the worst case scenario and I repeat, in the worst case scenario, shouldn't exceed (in terms of core(s) temp) a temperature of 89 to 94 degrees Celsius, which I find it hard to believe because my old processor (Intel® Core™2 Duo Processor E8400) TCASE is 72.4°C and when I stress it with prime95 the cores temp never exceeds 68-71°C, so why that processor cores temp are exceeding the TCASE temp? Is it the difference in the design between the two processors? Or is it a design fault in i7-4790k?

[RKimb]

ShadiElwan wrote:

The fourth information: If you tried everything to solve the problem, including updating the motherboard BIOS to the latest version and the problem continues, the processor should be replaced.

Unfortunately, my retailer says that my machine is "not designed for constant CPU heavy tasks", so I imagine I'd have a difficult time trying to get a replacement.

[SElwa]

Richard_Kimber, hi, I had the same problem with my retailer as you, but instead he told me something else, so I told him I will contact Intel and tell it the issue in detail, including what I did to try fix the overheating problem and if Intel tells me there is indeed a problem in my CPU, he would replace my CPU after seeing Intel's replay. So I advise you to do the same. Intel replied after 2 days, make sure you contact the warranty department as I did.

P.S. What does your retailer mean by your machine is "not designed for constant CPU heavy tasks"? Is that a joke?

[RKimb]

Thanks for the information.

I think I was dealing with someone fairly junior. Next time I'll email the Technical Manager. The PC was designed as a fanless system with a 95w fanless cooler. I've since added a couple of quiet fans. I'll see how it goes.

[cgeor]

Hello , i have intel i7 4790k and GIGABYTE GA-Z97X-UD3H , i have problem with voltage all in bios is defaut , but my voltage is 1.224 and my friend have same configuration and voltage is 1.164, what i can do to fix it ? please help