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.

[RBroc]

I changed my XTU settings to:

vcore: 1.090 v

turbo boost power max: 88w

processor current limit: 105 A turbo boost short power max: 105w

When stress testing in XTU I'm still hitting 85-90 degrees. In prime95 I'm getting 100+ degrees. When I'm playing Battlefield 4 my temperatures are 65-70 degrees. I find this still way too hot. Does anyone know what I need to change to reduce the temperatures?

Cooler: Neptune 240m watercooler Processor: i7-4790k Motherboard: Z97-G43

[DSzpa]

I too had the "instant temperature spike" problem with the 4790k. 100% CPU stress tests sent the temperature to 100 degrees C almost instantly. All the while my arctic CPU fan remained pretty much ice cold. Obviously the CPU line is faulty and not conducting heat to the internal heat sink properly.

No amount of Bios setting changes, fan reseating, buying new heatpaste would solve the issue. I urge anyone reading this thread: Do not waste your time listening to the Intel snakeoil salesmen in this thread. I too did so for 2 months, trying to fix the issue using bios settings, running the Intel Extreme Tuning Utility and changing the XTU variables. All that is a bunch of NON-SENSE. A smoke screen they use to distract you from the fact they have produced a faulty product.

I finally was lucky that a friend referring me to a local de-lidding service (as I have never done that before). I paid $30 and had my CPU back the same day. I put it back in, reseated the fan with the same heatgrease as before. Here are my results:

1. My idle (desktop) before was 42-43 degrees. Now 32.

2. Playing a 4k video before 75 degrees. now 45.

3. 100% CPU stress test with the Intel utility before 100 degrees. Now 62-64. See screenshot.

Anyone reading this thread with the 4790k heat issue you can do a simple test if de-lidding will help you. Open your case, run a stress test. Touch your (external) cpu heat sink. Does it feel cool/not warm? Then de-lidding will most likely solve your problems. Please not that you need to be aware of the risks of de-lidding. If you can find someone in your city who can do it for a small fee it might be worth it; it was for me.

It just angers me how Intel is responding to the problem and blowing smoke into people's eyes. I just wish some lawyer would see this and get some people together for a class action lawsuit. Intel most likely KNOWS their product is faulty but keeps lying to its customers. This is disgraceful.

[idata]

WOW!

I started reading bits of this thread a while back when, after changing from an antek all-in-one liquid cpu cooler to a component water cooling system, i noticed my I-7 4790K cpu was throttling at 100%. Of course i performed all the adjustments, tests, thermal compound replacement, throttle stop, is it my board, is it my chip, is it my cooling, yadda yadda yadda................ (really?) I returned to this "novel" after none of the above helped my situation. Oh sure i was able to bring temps down a little bit and mask the overheating problem with a software program called Throttle Stop, but i was flabbergasted when i continued reading and reading and reading all the posts from all the different CUSTOMERS, all with the SAME EXACT PROBLEM that i have.

Now, maybe I've been living under a rock lately, but i have NEVER read a more detailed, complex, enlightening and lengthy thread on one same issue as i have read in this one. I'm not even going to spout out my system component by component. Suffice to say we all have similar mobo's, coolers, programs, etc but the one thing we all have that is the same is an I-7 4790 (k). I have learned so much about overclocking, under clocking just from reading this one thread that on one hand i am thankful i found it. On the other hand i am very disappointed that there didn't seem to be a common resolution or "Best Answer" for, what appears to me to be a major, major issue regarding this particular CPU and the MOBO manufacturers. I am going to try some of the settings as suggested by many here, no wait, actually i am going to try everything that has been suggested here til i find a solution. But as one earlier disgruntled forum post put it, I have better things to do with my time than to have to be a diagnostic technician on a product i paid with my hard earned money, from a company that up until this point i trusted and relied on .

At the very least i would think there might be just a little more involved participation and response from the people at Intel (save Ken @ Intel thanks btw for your courageous efforts) to at least preserve their reputation as industry leaders in today's high tech world. There is a saying in sales that is backed up by research: A person who has a good experience with a company/product will maybe tell 4-5 people about it. A person who has had a bad experience with a product/company will tell 10-15 people on average. I know the high tech sales industry is a very lucrative one with companies relabeling obsolete overstock and reselling it as new etc.... competing for the consumers all mighty dollar. It's just a disappointing economic reality where the one who really loses out...... is the consumer.

[SKing4]

Same problem as everyone else.

I read this thread as well as others and tried all the advice within. I tinkered with all the settings in Intel Extreme Tuning Utility for hours and saw no results.

With Intel's stress test I'd get to ~95-99 within 2 minutes and shut it off there. I also used some heavy duty simulation/projection software that puts all cores to max load with similar temps.

Today I delidded my 4790k using the vice method that you can search for on youtube. I cleaned off the glue, stock thermal paste and applied Arctic MX4 and so far benchmarks are showing me ~10C in improvements. Intel stress test is maxing out at ~83, simulation software maxed out at ~89. Stock cooling.

I hear liquid metal cooling (liquid pro or liquid ultra) works much better than MX4 for this and people reporting ~5-10C lower using that, I don't have any available to me but plan to try that out later.

Also read a lot of reports of people with 4790k and similar processors having much greater success than this when delidding and applying new thermal paste (20C+ sometimes, which is what encouraged me to do this). Just youtube search "4790k delid"

Obviously you void your warranty but the amount of hassle it seems to take to convince intel to replace your chip, the warranty we have is kind of useless anyways.

[RBroc]

Somebody??

[idata]

I have the same problem.

Bought a Gigabyte GA-Z97X-UD3H, an i7 4790K and a be.quiet DarkRock Advanced cooler a few days ago, and I'm hitting >>90°C temps during stress tests as well, despite already undervolting the processor as far as I could without getting bluescreens. I'll try one more thing with getting really good thermal paste and reseating the cooler, if that doesn't work either, I'll see if I can RMA and get a correctly assembled sample as well.

Meh! Really hoped this build would run w/o problems, I need it for work.

[CPATA1]

Same problem with me. A few weeks ago I bought an Intel 4790K with an ASROCK Z79X Killer Retail.

Test were done with GSKILL TRIDENT X F3-2400C10D-16GTX @ 2400MHz, every test with prime95 v28.7 (small FFts, blend test), AIDA64 (FPU test), OCCT 4.4.1 failed with a blue screen followed by a restart. With XTU, values are set to unlimited . On a positive note, I managed to run prime95 v28.7 small FFTs test for 10 minutes @ default jedec specifications, meaning ram running @ 1333MHz.

It seems that CPU+MB can't keep up with high ram speed and particularly @ 2400MHz. Not to mention temps that overcame 80 Celsius degrees, with a NOCTUA NH-C14 and arctic cooling mx2.

There seems to be a problem with high ram speed. Everyone posting here pls also mention what ram are you using and with what speed.

I sent products back to where I bought them hopping for a refund. What a disappointment for Intel with such a product.

[BGray]

I bought my 4790k at the end of November, and put it together with a Z97S Krait Edition MSI board, G.Skill Sniper 1866MhZ ram, and a Cryorig H7 cooler. Before changing from stock to the H7 my bios temps showed mid 50s, and now they are about 40-43C.

So just changing to aftermarket cooling dropped the bios temps by about 12-15C, and maybe a 5-10C difference during gaming.

Prime95 (26.6) and Aida64 FPU was hitting over 90C within less than a minute. Running Asus RealBench I'm averaging about 68-73C. I've run the Intel Diagnostic tool and it always says I've passed, I've run the Intel ETU stress and maxed between 70-73, bench was maybe 75C max. Gaming I'm still running high 60s- low 70s, I've even tried adjusting the settings as recommended but that doesn't lower the temperatures any at all.

My idle speeds are around 29-35C, voltage at stock settings is about 1.168V, I've been told those are okay temperatures, but with so many conflicting reports on temperatures of this chip, I'm not sure what a good or even correct temperature with air cooling is. I'm guessing the jumping to almost 90C in the stress tests is not good, as I've seen others much lower than that with similar coolers.

With XMP on, and letting MSI turn on its turbo the voltage goes up to 1.23 and runs at 4.4 on all cores, and the temps go up about 5C more and the stress tests still run at 90C+, so I've had that

turned off for now, and just left my ram running at its correct speed.

My chip is a Vietnam chip, X530A662 batch, if that matters any.

Should I be happy with those temps, or should I keep looking into it?

I've done everything I could find, re-seated the cooler 4 times, bios is fully updated, tried different settings recommended through the ETU, no change.

[idata]

In the meantime, mine has gotten better (or is it just the cold winter air…?).

I changed the thermal paste to ThermalGrizzly Kryonaut but stayed with the be.quiet Dark Rock Advanced (and noticed my first thermal grease application had been spot on as well.).

Now I'm at about 26-33°C at idle and at 85-95°C when under full load in Prime95 28.x Small FFTs. And when using 26.6 I'm hovering around 65-75°C. So that's good enough for me right now, especially considering I'm turboing to 4.4GHz on all cores, even in 28.x, without any throttling.

Could it be better? Of course. But for now, I'm okay with it.

[PS11]

I thought I'd drop into the discussion too. I have bought a new i7 4790K about 6 months ago. Ever since I installed it, it displayed the same thing, very high temps of 85-100C at near 100% load, and instant spikes from idle temp(30C or so) to 90C in a second or two.

So far I have rarely had to run the CPU at 100%, but I'm starting to do this more often now, and it is getting to be a concern to me.

But the reason I am writing this post is because I have just installed a brand new water cooler for my CPU. My old one was a Gelid Tranquillo Rev.2, and with Gelid GC Extreme thermal paste. Applied it all correctly and I got above results. And right now I have bought myself a Kraken X61, one of the best cooling CPU water coolers. It had thermal paste pre-applied so I didn't touch it. Installed it all again, after giving the CPU a good clean with Isopropyl alcohol. I got exactly the same results. My computer tells me the CPU runs at ~95C right now, I am currently running folding@home. But, the Kraken app tells me the liquid temperature of the cooler is at 52C right now, and that's after having run folding@home for 2-3hours right now, with the CPU constantly giving temps of ~90C. The cooler fans are also running at only 25%, in a reasonably hot room, so it isn't venting loads of heat whilst taking in cool air.

I thought this would be a useful addition to the discussion.

oh, and a record for me, 100C just has been hit on one, no wait, two cores right now.

RMA? Right after I have installed the damned thing? Sigh.

UPDATE 1: Seems like the liquid temps have stabilized at 52C, with CPU temps drifting between 94-98C. GPU is at 64C. I'm new to liquid cooling, but how can it be that the CPU is at double the temperature of the cooling liquid?? Is the temperature sensor of the CPU faulty, or maybe the cheap thermal paste they put under the heat sink inside the CPU??

UPDATE 2:I'm now running at idle, temps of ~32C, and it shows both the CPU and the cooling liquid being at the same temperature of ~32C, so the temperature have a 50C difference at load, but at idle they are pretty much equal.

[FGarr2]

I´ve had an 4790 (non K) with the same issue for almost a year now (100°c at full load, thermal throttling and all). Intel and Gigabyte´s support didn´t help at all (Intel told me that even if the temperature goes above the announced maximun of 72/74°c, there's nothing to worry about if the computer doesn't fail, -I mean, how stupid can you think customers are, Intel?-), so I decided to take the risk and delid the processor to replace the internal thermal paste. It's not a difficult process, a bit scary at most, but it is to be done carefully as you can end up killing your processor. I did it through de vice method and it opened quite easily. I replaced both internal and external thermal pastes with coollaboratory liquid ultra (liquid metal), as it is regarded as the best thermal conductor in specialized forums. I also added a water cooler, but I think it's an overkill if you are not planning to overclock as the stock cooler should manage. The temps dropped 30 degrees at full load, and iddle is now at a stable 37° (we have 30°c right now in argentina).

I'm very sattisfied with the results, but would not recommend this process to unexperienced users as it voids your warranty and could end -very- badly.

In conclussion, Haswell processors clearly suffer from a design/manufacturing error and Intel doesn't seem to bother at all. Totally unacceptable as I was led to believe I had payed for the best quality product available. I'll think twice next time I'm buying PC components.

I found this forum useful when I found my processor was faulty, so I'm trying to give back some help to the ones who couldn't find/weren't given a solution yet. I hope it works out.

[FoxhoundUnit]

• Thank you everyone for the great informations . I'm trying to see how i can safely overclock my "Faulty" i7 4790K, I'm happy to find this old but still very accurate topic, With a Scythe Kotetsu Mark 2 cpu cooler i am able to turn Turbo Boost back on and get 81c Max (37 idle), And for further OC'ing like 4.5/4.6 i guess i will need to go water cooling.