Do you still want to buy an Intel CPU after reading this ???

Seanie280672

Estimable
Mar 19, 2017
1,958
1
2,960
Well it appears Intel are blaming users for the overheating issues on their CPUs and not the fact they use rubbish TIM and mount the IHS too high to the point where it doesnt make proper contact with the core, good old Intel.......they basically say "We do not recommend overclocking a processor that we charge more for the ability to overclock it."

http://wccftech.com/intel-i7-7700k-owners-flood-forums-with-overheating-complaints/
 

Seanie280672

Estimable
Mar 19, 2017
1,958
1
2,960


What they are basically saying here is, pay more for a K series CPU, but at the same time, if it has a "K", dont overclock it lol, I think they've lost the plot.
 
No, what they are saying is pay for an unlocked processor but dont come to us if you dont get the results you think you should be getting.
Intel (or any company) is not responsible for anything you do to their products, they can make it so you can do so, but that doesnt make them liable.

Even ignoring that, you pay the extra for the higher boost clocks over the non k processors, not that anybody ever mentions it.

That being said, this is what happens when you dont have a major competitor for several years.
 
No one officially suggests running their cpus outside of their stock specs. Ocing voids the standard warranty. This is just common talk even before this issue. If you want to try persuading people to stop using intel products, check out what they have done in the past but don't be biased and check what everyone has done. It's always been a pick your poison situation. Either that or don't live with technology.
 

Seanie280672

Estimable
Mar 19, 2017
1,958
1
2,960


Not being biased, ive used Intel CPU's for the last 10-15 years, the last AMD CPU I used before Ryzen was the Athlon XP, a lot of users on here wont even remember those CPUs without Googling them, and I also very nearly sent my RyZen back and exchanged it for a 7700k as I got fed up with waiting for AM4 boards in the UK, so glad I didnt, I had to de-lid my 2nd haswell after killing my 1st one due to temps, that was nearly 5 years ago, and still they've not learned.

This is just a warning, letting people know.

In June, Intel are releasing the 7740K, and if you look / red deep into it, it requires a new socket and addresses overheating issues, so lets face it, its a Kaby Lake refresh, same thing they did with Haswell, more money, zero performance gains, new socket (yet again), same speeds as the 7700k, but a higher TDP, minus the onboard GFX.
 

Eximo

Titan
Ambassador
Personally quite pleased to have a CPU running at 5Ghz. A little warm, but doesn't throttle.

My old Haswell also runs a little warm, so it runs at a modest 4.3Ghz. Still not complaining about 400-500Mhz for a little extra cash. Literally got what I paid for both times.

I have been thinking about de-lidding though as I have one core that always sets the bar. And if there isn't going to be another LGA1151 CPU I might as well.
 
I meant bias as when looking into the history of bad practices that companies have done.

7740k is not a kaby lake refresh (coffee lake), it's kaby lake. Enthusiast is always behind plus shouldn't be confused with mainstream offerings. That also means it has nothing to do with addressing mainstream issues. But if you get x299 with a quad core, I'd have to question your intelligence. Although there are certain situations that need it for more ram, they are uncommon. Still, enthusiast has nothing to do with mainstream so hwe is also not the same.
 
The only thing I noticed from intel's statement where they might have dropped the ball is not informing users of their performance tuning plans. They went outside the scope of normal use (stock configuration) and mentioned delidding in addition to overclocking, the 7th gen cpu's like past cpu's are covered under additional protection plans users can purchase.

For the 7700k it's $30 and offers the user a 1 time replacement of the cpu even if it was damaged from being run outside the factory specs. For what it's worth I'm not sure amd has such a program for users even though people have be overclocking their chips for years.
https://click.intel.com/tuningplan/

Their statement sounds like a standard coya comment. No company vouches for their product if it's run outside of the factory specs. While the k series are seen as overclocking 'capable', there's never been a guarantee of suitability in terms of clock speeds, voltages or temps beyond factory specs. All they've done is left the multiplier unlocked for people to tinker - at their own risk and with unknown results. It's uncharted territory and always has been. There's never been a k series cpu that says 'good for a 500mhz oc!'.

Some are better overclockers than others, it's part of the silicon lottery. People shouldn't be surprised by that. Just as people were complaining that their cooling fans were constantly spinning up and idling down. How is it intel's fault that users apparently don't understand how fan curves work? That's not something the cpu manufacturer should have to inform folks of considering it's controlled via the bios which is a motherboard situation.

It's not the first time I've seen fluff pieces done by wccftech or seen them quick to try and 'call out' intel. Where is their in depth article covering their experience overclocking kaby lake, or have they even bothered? Anandtech reported their findings along with plenty of details along the way such as thermals, core voltage, tests run. So has hardocp, including using a review chip, a plain off the shelf retail chip, performance with and without delidding.

http://www.anandtech.com/show/10968/the-intel-core-i7-7700k-91w-review-the-new-stock-performance-champion/11
https://www.hardocp.com/article/2017/01/26/intel_kaby_lake_i77700k_cpu_overclocking_followup/

Tom's put in the work.
http://www.tomshardware.com/reviews/kaby-lake-de-lidding-overclocking-test,4970.html

Looking for anything wccftech's actually done beyond whisper and regurgitate gossipy news from other sources still leaves me empty handed. Sort of like when they tried to turn it into 'bend gate' or something when a remark about the thin pcb's causing skylake and newer cpu's corners to bend because of 'heavy coolers'. Then of course if someone went far enough to track down the original story it was a review done by someone who admittedly used a power driver to tighten their heatsink to the cpu. A thinner pcb may have contributed to the issue but it didn't cover for plain old human error and was more about the mounting pressure which doesn't equate to heatsink weight. Even a waterblock (which is lighter than the stock cooler) can be over tightened and cause damage.

Fair to say wccftech enjoys 'juicy' news stories.

Similar to eximo, I'm running a devil's canyon i5 so essentially haswell. It has the funky tim instead of the soldered ihs that sandy bridge used. Coming up on my 3rd year of ownership, it's been oc'd to between 4.6 and 4.5ghz since day 1 depending on summer/winter (the room the pc is in often gets up to 32c+ in summer). No delidding, no custom water loop, just a large air cooler. Haven't even changed the thermal paste. It stays around 80-82c running p95 v26.6 and remains much lower during normal tasks and gaming.

Considering stock turbo only allows it to go up to around 3.7ghz with all 4 cores loaded I'm enjoying a solid 800-900mhz over stock. Not bad for the extra $20 or so I paid for a k series and that's running on the factory tim under the ihs. It doesn't mean the tim under the ihs is the best solution they could have used or that it might not be an issue for a few folks but in all fairness it's not that horrible.

Mine wasn't a cherry picked cpu from someplace that pretests chips and sells 'golden' cpu's or anything, just luck of the draw. Plenty of others with higher clocked cpu's than mine and with lower vcore. The idea that the factory tim is the worst thing ever and the downfall of desktop cpu's as we know them might be a little over dramatic.

Neither company is without their issues. Amd had trouble with the fx lineup and trying to sort out what actual core temps were, even their own overdrive software wasn't 100% accurate with the thermal algorithms. Those things still haven't been entirely addressed or there wouldn't be all the confusion over ryzen and why if they use the same basic 4 core clusters between the r7's and r5's, why is it only some of the r7's use a 20c offset? They don't use a different die than r5's and the 20c offset only applies to the 1700x and 1800x, but not the 1700. Reviewers are still scratching their heads trying to sort out whether those cpu's are or aren't overheating if and when the offset does or doesn't apply. Pick your poison.
 

biglizard

Respectable
Apr 1, 2016
314
3
1,860
Read that article about a week ago, bought my 7700k two days ago, installed it last night.

So glad I did.

4.7 at stock vcore 1.290.

73f Ambient.

Tonight ran Prime95 smal fft two hours max core temp 75c, IBT 10 passes max settings max core 74c and real bench two hours max core temp 65c. Couple hours of TW3 max temp I saw 54c.

Best advice I have for this chip is leave the cheapo AIO's where they belong, on the retailers shelf. Buy a Nh-D15 or equivalent air cooler.
 


So is that the cooler you have? NH-D15?? If not, what cooler do you have?
 

kgt1182

Reputable
Jun 8, 2016
420
0
5,160
I would seriously consider Cannonlake for Intel despite all their anti-competitive practices, ICC buas and monopolistic behaviour (stagnating performance).
The performance figures of Cannonlake 10nm are very convincing. Even GloFo's "10nm" compared to their own "14nm" yielded the Exynos having a 43% power consumption drop with same power consumption, allowing a 20% clock boost if perfected (HiSilicon also did this, its like 22nm to 14nm) Granted those are small chips, and Intel predicts a >15% performance gain, I will not be surprised. We have the 22nm 3770K vs 14nm 6700K/7700K, and the 8700K even at the same IPC could easily hit 5.0 GHz base assuming a meagre 10% clock gain from scaling.

That is after KabyLakeX/ Skylake X, Purley and their US$12400 Xeon Platinum 8180 (US$16200 for MCDRAM 8180M if you prefer)

That being said, we may have more than Skylake IPC with clocks more than the most extreme of Piledriver FX 9590 phase change overclocks on air at stock.

And they will still use thermal paste for heatspreader, I will delid, you should too. I will buy Intel if the performance scales well.
 

biglizard

Respectable
Apr 1, 2016
314
3
1,860


NH-D15S.
 

razamatraz

Honorable
Feb 12, 2014
200
0
10,710
I can't disagree with tossing some flak Intel's way over this but all those new processors are for the HEDT platform; they are not refreshes to the mainstream platform. We may well still get Coffee Lake on 1151/Z270.
 

Astralv

Distinguished
I don't know what they talking about. I have Ivy Bridge, runs about 37C on stock cooler, Hesswell 4770K- runs Hot- about 47C on stock cooler and Kaby Lake 7770K- runs about 32C on basic Evo air cooler. Stop looking at your temperature. I did not check my temperature for 3 years, see it occasionally if I get in to BIOS. Also I did not see any issues with Kaby.
 

Eximo

Titan
Ambassador
Not to point out that those are likely idle temperatures from a cold boot, but a small update:

Delidded my i7-7700k and applied high end thermal compound, not liquid metal. Temperatures went from low 80s, with one core consistently hitting 87-91, down to a high under bench of 69.

I probably won't do it to my old i7-4770k, since it is getting passed along to my brother for his graphic design work. Though that one was pretty well fine at 4.3Ghz, even with an Evo 212 on it it barely broke into the 70s.
 

Max_x2

Distinguished
Mar 26, 2013
173
0
18,710


So you managed to kill an haswell by overclocking? Wow, I honestly thought it was near impossible. There isn't much I haven't done to my 4670k since I bought it on release, and it still runs like a champ.
 

Seanie280672

Estimable
Mar 19, 2017
1,958
1
2,960


Nope I killed the 1st one's IMC after de-lidding it so threw it away, brought a new one and de-lidded that one too, 2nd time successful, I see the confusion, de-lidded it them due to temps.

What makes you think its near impossible to kill any chip ? accidentally put too much voltage through, like type in the wrong numbers without thinking about it, save and exit the bios - BANG !!! there goes your chip. :hot:
 

Max_x2

Distinguished
Mar 26, 2013
173
0
18,710


I've put mine to something like 1.5V with 2.1V input, and ran like that for couple days. Stressed tested and all. I was prepared for the fact that it could degrade/burn. And an "accident", man... I usually go through my settings 3,4 maybe 5 times before pressing f10 when i'm overclocking. Especially if it's an "extreme" o/c. Plus, usually, the chip will protect itself and either not "start", or shutdown real quick.

I know it can happen... Zall i'm saying is that you have to be either extremely sloppy, or extremely badlucked.