Gamers Red Hot with Fury over Intel Core i7-7700 Temperature Spikes

The problem is that they charge more for the K cpus. As a matter of law, the buyer pays for something. Since Intel has links on the site to overclocking reviews, there is no question what those customers are buying. For them to somehow state they aren't selling those cpus to guys who will overclock is disingenuous.

Please show me where they charge more.
 
As we reported on yesterday, some users are getting increasingly frustrated with their high temperature spikes on their Intel i7-7700K CPU's, while Intel has just shrugged at the matter stating that it is normal. Well, users are getting increasingly flabbergasted at the details of Intel's response: If you don't like the heat, don't overclock your CPU.

Personally I am torn on this one. On the one had it is clear that something is different with these chips than with other recent Intel chips. Hitting temperature spikes of 90°C at stock clocks and voltages is clearly not normal, and I would clearly be disappointed if I got a chip with this problem, and couldn't overclock it.

On the other hand, overclocking has always been a "your mileage may vary" type effort. CPU manufacturers have only ever guaranteed that they would work at stock clocks. So, while I might be disappointed, I'd probably just call it "losing the silicon lottery" and deal with it. Some are suggesting the fact that Intel sells a K variant means that they should stand behind overclocking, but I'm not quite sure I agree. A K variant means "we won't stop you from trying", not "we guarantee you will succeed".

Thanks to SpeedyVV for the heads up.

We do not recommend running outside the processor specifications, such as by exceeding processor frequency or voltage specifications, or removing of the integrated heat spreader (sometimes called “de-lidding”). These actions will void the processor warranty.

You do know that AMD says the same among others? And do we even have to talk about the overclockers dream chips including those with the 4Ghz wall.?

And where is the 90C documentation?
 
90 Celsius is not normal at stock speeds for any chip, let alone one they market to people to overclock. That horrible TIM they use on top of a possible voltage issue is a design flaw and would piss me off if I owned one. The answer from Intel was weak and a massive disservice to people that support them, basically told them to F-off. I hope they get hit with a class action lawsuit, they need their hand slapped and a reminder pissing off customers is a bad thing.

You be surprised. What do you think the real temperature of Ryzen and FX/APU chips are if it was measured the same way instead?

And again, 90C is WAY within specs even if you want to go that route.
 
I sent the heads up to this article.

This "issue" to me is actually a non issue, what I really thought was interesting was the bullshit response.

You really gotta love these marketing/PR type responses.
 
A friend recently got a new PC from newsegg, 7700k and Asus board, 32Gig etc.. nice setup.

He had problems with stutter in games he couldnt solve so I dialed in, did a few tests with Aida64 and prime95 with and w/o AVX code.

Well, either the guy who put it together made a mistake or this CPU is REALLY at the edge of specs..and beyond...as it reaches 102°C in a few seconds under any prime version and Aida64 stability test fails in less than 1 minute, temps also concerningly high.

In the end we found out that his stutter is a TrackIR version problem which we could easily solved but found this temp issue to be real critical. It will run some games at up to 80°C without errors now, thatÄs all he actually wants, but the system remains unstable

if you consider that it heats up that QUICK and triggers the limiter tp protect itself in all stock settings, newest Asus bios and all default, no XMP, nothing. OK, he "only" got good air cooling, but can it be that you need high end watercooling to run that chip under load ?

I am not sure ( havent tested it ) if that machine would do a real world handbreak recode if he was about to work on his video library !?


Baseline is, that guy is 74 years old, lives half way around the globe and Teamviewer sadly doesnt allow to repaste a CPU/Cooler combo yet, he is scared of doing so and also is not to happy about sending it back either, which was my advice. I said send it back and get it fixed as it doesnt hold any testrun, not any !


He still runs it, flies his DCS flightsim and feels happy,, just I have a bad feeling that something is wrong, either a real bad CPU ( worth an RMA ) or the cooler/paste thing got fucked up by who ever assembled it.


This "news" makes me tend to believe that his special CPU is just a bit beyond Intel's "just spiking to 90°C and its OK" theory.


Time will tell if more reports surface about 7700k just not staying at 90°C but 102°C and more till shut off.
 
I sent the heads up to this article.
This "issue" to me is actually a non issue, what I really thought was interesting was the bullshit response.
You really gotta love these marketing/PR type responses.

Did you expect anything else from a company that already cut down costs on lids and conductive paste ?
You do know that AMD says the same among others? And do we even have to talk about the overclockers dream chips including those with the 4Ghz wall.?
And where is the 90C documentation?

Nice way to spin this one. Maybe just blame AMD for everything , takes much less words then what you are typing here ?
 
I guess I won the silicon lottery. I've had nice stable 30C idle temps at 5.0ghz with no spikes, 80's under avx 100% load and 70's for 100% load anything else. I'd go crazy if I hear it spike up the cooling fans over and over.
 
Has there been any common thread like a certain motherboard used or something else which may precipitate this specific issue?
 
Its time to ditch intel and go with amd. Pretty pathetic on their response. Lots of loyal intel folks here and i think folks should start migrating to amd. As for me 3770k will be my last intel chips i will ever buy from them. Suck you intel
 
So how do you define "running out of spec"?

I define running out of spec as running the voltage or frequency as higher than set by default on the chip. Intel says not to do that, therefor I conclude they are saying "don't overclock".
 
The very term "Overclock" pretty well defines that you are operating the device outside of the normal specifications whether the designer makes it easy to do or not. Not much different then buying a Ford Mustang and changing the exhaust, intake, injection system, adding a larger turbo, and reprogramming the computer to get several hundred extra horsepower. Just don't expect Ford to honor the power train warranty when your #4 piston departs at high speed.

Now the really worrisome thing is that several articles on this issue are stating that normal CPUs operating at specified clock speeds are also showing the problem.
 
What about when XMP sets your otherwise stock system to a figure above stock? I'm running manufacturer specs and it clocks my CPU at 4.5 Ghz, this isn't a desired effect for me currently.

So if I weren't knowledgable in fixing this myself, would the idiot user who 'followed directions' have his warranty voided too?
 
What about when XMP sets your otherwise stock system to a figure above stock? I'm running manufacturer specs and it clocks my CPU at 4.5 Ghz, this isn't a desired effect for me currently.

So if I weren't knowledgable in fixing this myself, would the idiot user who 'followed directions' have his warranty voided too?

There is a certified list of that at Intel. CPU+mobo+memory.
http://www.intel.com/content/www/us/en/gaming/extreme-memory-profile-xmp.html

Mobo makers auto overclocking however in strict legal terms void warranty.
 
Maybe you should learn to check what you link before posting it.
Boy, you are an asshole, aren't you?

Do they charge more? Do they? You were denying it a few posts ago. Now you are deflecting and attacking the poster. I have seen what others have said about you but since I don't follow every post I have been giving you the benefit of doubt. I guess they are right.
 
You are making a bunch of weird statements. 1st, if you really did your homework, you would know that on the page where Intel has posted the cpu specs, they have links to over locking reviews. Then they make a statement that you should not overclock beyond specs and they specifically mention frequency. Look at the specs and tell me what the frequencies are and then compare to the articles linked. Now take that to court and see who is right.

Let me try and make this simple for you.

The CPUs in question in the thread work, they even work while being modded to the extreme with heavy overclocks (most of them auto overclocked as posted already), the people are complaining about a temp spike that they just dont like, still within thermal limits of the chip and not killing the chip. Even if Intel said they warranty ALL chips against failure from OCing, Intel would not warranty the chips for not holding a desired temperature range. That is what these people, and yourself are complaining about. Intel never said not to OC, just like a car company might sell a sports car, and even pay and support race teams and post race results for their team etc etc, that does NOT mean you can take your Miata to the track, have it over heat and bring it in for warranty work. Now, if they know that or not and service the car or not is something else.

People like you are why we can't have nice things, or have companies involved in the community, because if they are, or as you say, simply LINK to some writing about overclocking, that must mean everything is acceptable. That is just childish logic. If you don't like the terms, don't buy the CPU. But as also posted already AMD states the same thing, but does not offer the option of the tuning plan that Intel does.

Have you read the pages of posts on the forum? About half way in, someone comes along who knows some OCing and finds out most people are still using SVID and auto overclocking, as already mentioned here a few times, these people found their voltages going to 1.4-1.5V! High voltage being common with auto overclocks, the guy walks MANY of them through setting manual voltages and magically the people posting after "WOW! Huge drop in temps!!!", in other words, lots of people who have NO idea what they are doing, but getting some HW and setting a high frequency and letting the mobo do everything else.

Even the main talker there admitted after doing this, his spikes went away. But being a talker that was not good enough, then he comes back pages on to complain he can't hit as high of a stable OC with manual voltages, mostly because he doesn't know what he is doing and his cooling solution can't keep up and he doesn't like anything over 80C. So rather than learning to OC better or getting a better cooler he goes back to bashing and saying Intel needs to do something about his own incompetence.

So how do you define "running out of spec"?

I define running out of spec as running the voltage or frequency as higher than set by default on the chip. Intel says not to do that, therefor I conclude they are saying "don't overclock".

They have 3 documents they list spec and limits, for the 7700k the voltage limit is 1.52v, no frequency limit given. They said out of spec, not out of base/VID etc.

Voltage limit and other key factors/limits are listed in Volume 1.
 
Last edited by a moderator:
The very term "Overclock" pretty well defines that you are operating the device outside of the normal specifications whether the designer makes it easy to do or not. Not much different then buying a Ford Mustang and changing the exhaust, intake, injection system, adding a larger turbo, and reprogramming the computer to get several hundred extra horsepower. Just don't expect Ford to honor the power train warranty when your #4 piston departs at high speed.

Now the really worrisome thing is that several articles on this issue are stating that normal CPUs operating at specified clock speeds are also showing the problem.

Bad analogy. Your example is equivalent to Intel denying warranty because you delidded. You've modified the product, they no longer have to gaurantee the part.

This temp thing is more like Ford denying you warranty because you drove your truck over the speed limit and up to the point where the governor kicks in and THEN it blew a piston.

That is to say, you were doing something unusual and not recommended, but the product would have never failed if it wasn't defective in the first place.
 
The reality is this is probably a non-issue. Are chips dying or throttling as a result? This may manifest itself as early death of these CPUs however I don't think we are there yet.
 
I'm starting to lean on the auto-overclock of some motherboards and incompetence of some users for this mess after delving a bit into the other sites and threads. On a work dev machine I built awhile ago, I noticed fan cooling instability of a asus z170 tuf motherboard and 6950 combo when it was in auto-overclock mode.

It was screwing up my machine learning calculations on the "hello world" equivalents and I was getting frustrated by weird random locks happening. My EVGA Z270 classified doesn't auto overclock, I did an initial windows 10 install, software install, and windows updates at stock and never saw speeds outside its normal boost nor did I hear my corsair 115 start blitzing its cooling. Properly configured at 5.0 ghz, it has been rock solid and silent with no spikes in temps.
 
So how do you define "running out of spec"?

I define running out of spec as running the voltage or frequency as higher than set by default on the chip. Intel says not to do that, therefor I conclude they are saying "don't overclock".

I should also state, that even if Intel only allowed what you are suggesting, that is their choice. And according to what people posting here think, that would mean if Intel will not cover all OCing issues, that Intel should not offer an unlocked chip, which beings me back to "this is why we can't have nice things".
 
I'm pretty sure when I bought my 2500K, overclocking was a key feature and the K CPU were the only Intel CPUs covered if it broke because of overclocking...
I'm remember they specifically gave a warranty on the K part for defect during overclock.

I might be wrong so feel free to link me to the truth :) (Please no bashing for no reason, I'm merely trying to recall my memory here.. maybe I should google first hehe)

EDIT: Apparently they sell a protection plan on top of the 3y warranty for that at ~30$ see below url
https://click.intel.com/tuningplan/purchase-a-plan
https://www-ssl.intel.com/content/www/us/en/gaming/overclocking-intel-processors.html

I wonder if it was like that 5y ago too.
lol CPU DLC :D
 
Lemon or not to Lemon?

If I sold you a car that automatically controlled cabin air temperature and when you moved from a shady spot to a sunny spot on a cool day the A/C came on at full blast for 2-3 seconds EVERYTIME without fail..... you might call it a lemon.

Sure, it might be a 0 overclock cpu because they already turbo boost near their max and you got a lower max clock chip, and maybe the chip manufacturer will lay the blame on the end user for modifying their product and swipe the issue off their desk...

You cannot escape the fact that numerous people have reported their cooling systems going from 20% fan speed to 100% because of the thermal spikes.

I personally dogged the whole issue by going with a 6850K and it runs beautifully at 5.0GHZ, but I know people who believe they got screwed on choosing the 7700K and are going the RMA route until they get a working one path.

Just my 2 cents, but if you sell a product that impacts the user experience and pass it off as a feature, you might have damaged your rep.
 
I suspect more and more folks will be seriously looking over those Ryzen cpus and mobos :D

Really not a good time for Intel to get political rather than investigating the issue in depth.

Right now I'm really glad I held off ordering that 7700K ;)
 
While Intel is surely fine legally here, the spirit of the k skus was to overclock. No guarantees, but go ahead and try it. That said, a ryzen 1800x doesn't stack up too poorly to a non-5ghz 7700k
 
I've had no problems with my 7700K, and it's running at 5Ghz on air. Haven't seen any significant heat spikes.
Spent some time testing to get the voltage as low as possible while still hitting 5Ghz, but it's been running great once I got all the setting tweeked.
 
Let me try and make this simple for you.

The CPUs in question in the thread work, they even work while being modded to the extreme with heavy overclocks (most of them auto overclocked as posted already), the people are complaining about a temp spike that they just dont like, still within thermal limits of the chip and not killing the chip. Even if Intel said they warranty ALL chips against failure from OCing, Intel would not warranty the chips for not holding a desired temperature range. That is what these people, and yourself are complaining about. Intel never said not to OC, just like a car company might sell a sports car, and even pay and support race teams and post race results for their team etc etc, that does NOT mean you can take your Miata to the track, have it over heat and bring it in for warranty work. Now, if they know that or not and service the car or not is something else.

People like you are why we can't have nice things, or have companies involved in the community, because if they are, or as you say, simply LINK to some writing about overclocking, that must mean everything is acceptable. That is just childish logic. If you don't like the terms, don't buy the CPU. But as also posted already AMD states the same thing, but does not offer the option of the tuning plan that Intel does.

Have you read the pages of posts on the forum? About half way in, someone comes along who knows some OCing and finds out most people are still using SVID and auto overclocking, as already mentioned here a few times, these people found their voltages going to 1.4-1.5V! High voltage being common with auto overclocks, the guy walks MANY of them through setting manual voltages and magically the people posting after "WOW! Huge drop in temps!!!", in other words, lots of people who have NO idea what they are doing, but getting some HW and setting a high frequency and letting the mobo do everything else.

Even the main shit talker there admitted after doing this, his spikes went away. But being a shit talker that was not good enough, then he comes back pages on to complain he can't hit as high of a stable OC with manual voltages, mostly because he doesn't know what he is doing and his cooling solution can't keep up and he doesn't like anything over 80C. So rather than learning to OC better or getting a better cooler he goes back to bashing and saying Intel needs to do something about his own incompetence.



They have 3 documents they list spec and limits, for the 7700k the voltage limit is 1.52v, no frequency limit given. They said out of spec, not out of base/VID etc.

Voltage limit and other key factors/limits are listed in Volume 1.
I have a problem with their statement. They mentioned frequency there.
 
I suspect more and more folks will be seriously looking over those Ryzen cpus and mobos :D

Really not a good time for Intel to get political rather than investigating the issue in depth.

Right now I'm really glad I held off ordering that 7700K ;)

I don't see your point at all except in theory at best. Again, there is nothing to investigate. The CPUs work fine at stock, and not even the K moniker means it WILL overclock, just that they won't stop you from trying in the easiest ways possible. Expecting more than that is unfair and unreasonable. And enough users at both stock and overclocked have chimed to report no such spikes so it's NOT a chip wide issue. Intel won't waste time and nobody but AMD fans who wouldn't have bought a 7700K to begin with will it make a difference for.

What you forget is that MOST K buyers DON'T overclock! They buy the K model for the higher turbo frequency, that's it. MANY K machines out there never even have the user looking at monitoring software! Most, in fact. Mountains out of molehills, man, mountains out of molehills.
 
And the topic, none of them have had a failed chip, their issue is they have a temp they don't like on a CPU that's working just fine, even outside of spec and being delidded, and Intel stated that they can not warranty that.

So now that you have nothing to go on, you just start making up scenarios?

As for getting into a dispute, I actually read the terms first unlike some people it seems. Which is why both of my CPUs have the Intel tuning plan on them.

I said 3 posts ago that you are either incapable or unwilling to understand the point I'm making, and you continue to be obtuse.

  • I *never* applied my statement directly to this issue. It was always a general statement.
  • I didn't mention temps running high.
  • I didn't even say I was going to RMA my own disappointing, yet functional chip.

My point, all along, is that they can't deny warranty simply because you OC'ed your CPU. Their "terms" are not ironclad simply because they say they are. Also, congrats on blowing $60 on insurance you don't even need.

People like you are why we can't have nice things, or have companies involved in the community

I repeat what I said earlier as well: "Pfft. Tool". Intel is a publicly traded company, with fiduciary duty to its shareholders alone. Anything they do that doesn't enrich their shareholders is actually a breach of duty by the CEO and the appointed officers of th company. Anything Intel or any other public company does for the community is to generate goodwill (which is an asset) via positive press. No public company helps you out of the goodness of their hearts or the size of their margins. Talk about childish...

Nor do they need you defending them for free. At what point are you going to stop giving these people free money and your free time?
 
Last edited:
Intel feeling the heat from Team Red

Poor Intel...poor poor Intel
 
This "Principal Engineer (I Do not Represent my Employer on Twitter)" at Intel doesn't seem to have a problem of overclocking the 7700K.. especially when it is done with LN2..

frankie.png

He is the Intel "principal engineer" who had a problem with Kyle's 7700K overclocking attempts back in January and who famously tweeted..

frankie2.png


Apparently he just recently learned a valuable business lesson about marketing..

frankie3.png
 
Back
Top