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.

[MCron]

I am also seeing abnormally high load temperatures.

ASUS Maximus VII Impact

i7-4970k

L4 batch

Stock cooling fan applied correctly. I have even taken it off, verified the paste applied, and reseated correctly.

I bought the CPU last week. As soon as I try video encoding temps go from 40C at idle up to 100C and the CPU thermal throttles. If I leave the system at this for a few minutes the system freezes and reboots.

This is at stock speed and stock voltage. Stock BIOS settings. Case is getting plenty of air and I have even run with the mobo sitting alone by itself in a 25C room temp.

If I disable turbo the temps max out at 85C and the CPU never thermal throttles. I'm running with turbo off for now. System is fully stable, but still too warm for stock speed.

@Intel, this seems like a bad batch of CPUs. I'm fixing to return the CPU but it seems like a new CPU is not the answer.

[YYuri2]

I have build me a new rig yesterday using the i7 4790K and the Asus Maximus VII Ranger motherboard.

And I aswell have the same overheating issues, a short google search made me end up in this topic.

I played some BF4 to test and the CPU started to heat up to 80 degrees within the first minute...

As Richard_Kimber already said, the reported temperatures go up and down very very quickly. The very same thing is happening here.

Is intel still researching this issue and when can we expect some more info regarding this matter?

[YYuri2]

So everyone responding in this topic with the issues have a L4 batch CPU?

Intel is too quiet at the moment imo so I guess we have to do our own research..

[RKimb]

Mine was described by the retailer as:-

Intel 4th Gen Core i7 4790K 4.0GHz 88W HD4600 8MB Quad Core CPU

[YYuri2]

And does the sticker on your box says: Batch: L4xxxxxx ?

[RKimb]

Actually, I've discovered that with certain Prime Grid tasks, temperatures can go up to 100C even when Boinc is set to use only 10% of CPU time. I think I'll stop using Prime Grid until this issue is sorted out.

[RKimb]

They didn't send me the CPU box. It was a PC that they built.

[AStar11]

Hi. Just checked my processor box:

Batch L4 indeed, however my chip is working correctly now, with H100i never exceeds 60C, even under 12 hour stress tests @ 4.4 Ghz. I have it undervolted to 1.16V. So either not all chips are affected in this batch, OR it's not the actual processor that's the problem. The latter is my theory. I think it is something to do with the microcode for certain chips and voltages, maybe? Flawed packaging is also a possibility but it doesn't affect all L4s since mine works fine now.

Also Do Not run Prime95 v28. I have tested with V26, Intel Burn Test, and AIDA64 Extreme and my temps never exceed 60 @ 1.16V, full 4.4 on all four cores, eight threads.

One thing to definitely try is make sure the voltage is not on Auto or Adaptive and manually specify it. I have noticed that CPU-Z sometimes incorrectly reports CPU voltage. So use software that came with your motherboard if you have any to read it. It should not be much higher than 1.21V, really. Also, Haswell processors use a "input voltage" or "VCCIN" which goes from the motherboard VRMs into the Integrated voltage regulator on the package. This is Auto at 1.8V for me, I have also lowered this MANUALLY to 1.7V. Might want to increase your load line calibration (or similar) up a notch, since mine was dropping to ~1.6 under load, but now really sits just below 1.69V and is 12 hour stable. This also helps to reduce temps on the package. I leave CPU cache voltage where it is.

Try manually lowering the voltage to 1.2V for the CPU VCORE, or even lower, mine is 24/7 stable at 1.14V, but I run 1.16V for extra stability. This should ensure that the motherboard isn't doing anything funky with the voltage. Apart from this, I'm not really sure what else I did after re-seating the cooler several times.

Hope this helps, and good luck.

[JBras1]

I'm also having a cpu from batch 4, my guess is that this batch has a lot of faulty CPU's in it.

[ATama2]

Yes ,

I checked , it is

Batch# : L421C001

[MDitz]

I hope Intel will come with some sort of statement soon. I am on the verge of buying the 4790K, but if there is a bad batch, or there is a design flaw I will not get it. But I first need that information. If I buy I will of course not use the stock cooler, cause they are really bad, so I have never used them anyhow. But I hope for myself and everybody on here that Intel will write information soon.

[JBras1]

I don't know the exact percentages of people with this issue and the people that are aware of having them. Right now I'm wishing I didn't buy this CPU because I'm at a loose end here, I can't fully use my new rig right now and it's still unknown what the exact issue is here. This issue has been reported 2 or 3 months ago on these forums and yet Intel has not given us any detailed information or statement about this problem yet, that's really kind of worrying. I just want to know what the next step is I have to do. I already send my first CPU back and now I got a new CPU that just gives me slightly lower temperatures, do I have to send this CPU back as well? This motherboard has been tested now and works just fine with a different CPU-type, so I guess it's either a combination of both or a bad CPU.

I would not recommend anyone buying this CPU yet as a lot of people are having issues with overheating, I'd simply not risk it. I hope more people with this issue will respond here and hopefully that'll trigger Intel to help us out with a real solution to this problem!

I have tried messing around with the settings Sponge gave us, lowering the voltages gives me some small beneficial results but I can't get the same results you are getting though. Decreasing the Vcore settings seem to give me the most beneficial results as of yet, I will waste another few hours on this problem and I hope I can get at least something acceptable.

[YYuri2]

So I got fed up with the high temperatures and waiting for a response from Intel and decided to try a new aftermarket cooler; Scythe Mugen 4

Results were quite stunning as the temperatures dropped by 30 degrees, and won't get hotter than 50 degrees under full stress!

To compare, the temperatures were 80-85 degrees while playing a regular game before.

Currently I'm still running with all the voltage settings on auto and boost enabled.

[AJash1]

I also have the same problem with overheating i7-4790K.

When idle the temps are between 25-30 C (with a new CPU-cooler, not stock). When the load is about 50% the temps within seconds rises up to 50C. With the stock cooler it went up to 70C.

I looked at the batch and it is indeed L4.

I use MSI Z97 gaming 5 motherboard and recently updated to the latest bios, without improvement.

I hope we get any answer from Intel soon.

[JBras1]

I have been playing around with the settings and I think I have fixed the problem. I will test around some more and will get back to you if I know the exact problem.

It seems that my motherboard was boosting my CPU to run on abnormal settings automatically and manually setting these to stock settings will help you out.

Using Intel Burntest gives me around 57c maximum and that's the way it should be.

I'll explain you what I did so people can try the same. I have used some settings people described earlier and I'll try to describe this as easy as possible so even the non-technical out there can solve this. Keep in mind that that I'm using a Gigabyte Z97-X Gaming 5 Motherboard and the BIOS interface might be different depending on what kind of motherboard you are using. I hope this helps some of you out and feel free to contact me if you are still having issues.

1. I've manually set a CPU Frequency of 4.00GHz in the Advanced CPU Core settings, when my motherboard was boosting it up to 4.4GHz things get very hot  

2. Set a powerlimiter of around ~88watts manually in the Advanced CPU Core settings, the Intel Turbo Boost was cranking up my wattage to abnormal levels as soon as you start stressing your CPU

3. Set a Core Limit of ~105 amps in the Advanced CPU Core settings, same deal

4. Disable K OC in the Advanced CPU Core settings (I think this setting only applies to Gigabyte motherboards)

5. I've set the Vcore around 1.19V

These settings at least fixed the issue for me, please let me know if this helped you out. I think with some tuning here and there you can get your own desired preferences.

[RKimb]

I tried your suggestions. Thanks. On my machine the CPU frequency was fixed at 4.00GHz and it wasn't obvious that it could be changed anyway. Similarly the power limit was already set for 88w, but I disabled Intel Turbo Boost. The Core Limit was already set for 105, and I disabled the K OC. I set vcore to 1.19.

This seemed to make a small difference, but I don't regard the result as satisfactory. If I now run Prime grid using 45 % of CPU time I get temperatures varying between 54C and 92C, with temperatures in the 80sC for quite a lot of the time. I don't think it's a solution for me. I'd expect to be able to run Boinc at 75% of CPU time without problems.

[JBras1]

I'm sorry to hear that you are still having problems, I think you have it even worse then me. My CPU started overheating when I was around 70-100% CPU-usage, looks like it's already happening for you at 50%. Can you try and check how much wattage your CPU is taking once you are experiencing the overheating issues? Playing around with the powerlimiter and corelimiter gives me very different results, from 60 to 100 degrees at 100% load.

[RKimb]

I'm not sure I know how to check that while I'm running the program. The power is already limited to 88w. There comes a point in all this where reducing the speed of the computer somewhat negates the reason for buying it in the first place.

[RKimb]

I now see that at the top of this thread it says that this question has been answered. I must have missed something. Can someone tell me what the answer was? Have the Intel engineers produced a solution?

[Allan_J_Intel1]

To all:

The L4 does not indicate a batch. The "4" in the batch simply indicates the manufacturing year. In this case, 2014, since we just released this product all CPU's will have a manufacturing year of 2014.

Once I get more information about this issue, I will update this thread.

Allan.

[REnso1]

The selected answer should be quoted beneath the OP, it is in my browser. Otherwise it is /message/253580# 253580 here.

It isnt really an answer, more the summary of how far I was able to get with the advice of others including my retailers after all the testing and replacing the CPU. I thought by selecting it as the answer it would make it easier for people with similar problems to get to the relevant facts quicker rather than read the whole thread through.

I did not intend selecting a correct answer to close down the discussion. I have found it very interesting to read what people have added to the thread and hope we can get some advice from Intel engineers about questions raised by other contributors, I understood these had been passed on to a different BB so I hope answers will still be forthcoming, as looking around the net at other tech BBs there are a lot of people struggling with this problem of overheating 4790Ks and all could do with some direction.

The impression I am getting is that the I7 haswells are hot CPUs. With my case closed up and a couple of more powerful outflow case fans http://www.akasa.co.uk/update.php?tpl=product/product.detail.tpl&no=181&type=Fans&type_sub=Ultra%20Quiet%20Fan&model=AK-FN073 140mm http://www.akasa.co.uk/update.php?tpl=product/product.detail.tpl&no=181&type=Fans&type_sub=Ultra%20Quiet%20Fan&model=AK-FN059 120mm added my replacement has been running at mid 70's in games with 44x4 turbo and I have had to bring it down to 41x4, 42x3, 43x2, 44x1 to keep temps in the 60s (ie below http://ark.intel.com/products/80807/Intel-Core-i7-4790K-Processor-8M-Cache-up-to-4_40-GHz Tcase, updated value 74.04°C) with a Noctua NH-L12. To get it to run 4 cores at 4.4GHz turbo I would have to get a much better cooler. Unfortunately the case I have is too small and was designed for air before hydro AIOs bcame so popular, so all the fans are in the wrong places for rads. I could get a new case only it is a gift from family so it has sentimental value so I will accept the reduction in top turbo speed for that reason. My other option is to delid it and sort out the TIM thickness but I really dont feel like it at the moment having just put it all together.