![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Gamers Red Hot with Fury over Intel Core i7-7700 Temperature Spikes
- Thread starter Megalith
- Start date
D
Deleted member 243478
Guest
LOL the 6700K's have been doing this for ages, they just don't hit 90c... more like 77c
Its a thermal interface issue with the IHS as it spikes instantly before heat can be transferred to the cooler.
Its a thermal interface issue with the IHS as it spikes instantly before heat can be transferred to the cooler.
It's the degradation from extreme heat people are concerned with the most I believe 90c is pretty damn hot, but not unheard of either I think my old 260GTX GPU use to run at around that however look how old that is by comparison and smaller node chips I think are even more sensitive to heat tolerance at such high temps.
dgz
Supreme [H]ardness
- Joined
- Feb 15, 2010
- Messages
- 5,838
Most people are not affected therefore there is no problem. We should apply this logic to everything. It's wonderful
Broadly, there is no problem with Intel. The problem seems limited to a small enough segment to need to look to other causes then the chip itself. Simple as that. Unless there's a bad batch, then the chip is not the issue here, clearly. There is no evidence currently to support any supposition of a bad batch yet. Should such come to light, we will revisit this as a "Intel issue". Until then, it's an motherboard level issue or user issue or Intel would be more worried about it. The logic is pretty simple and straight forward.
I followed the old rule I've used on many other CPU's over the years.
I set the voltage close to the max setting and see the max stable speed for basic tests. Max was 5.2 Ghz on this CPU
Then I lower the voltage and see how low I can go while keeping it stable (and of course to reduce the heat).
Then I up the voltage a little or lower the speed a little to get it stable in heavy stuff like Prime95 or HD video compression.
There are other setting that affect the voltage under light or heavy loads on my board. Voltage varies from .982 (no load) to 1.36 (full load)
Also, the AVX offset in my sig setting drops it to 4.3Ghz when running Prime 95 or some video compression apps.
The offset is more to keep the CPU under 85c than a stability issue. I could run the fans at a higher speed, or switch to water cooling, but it's not worth the extra noise/costs for an extra 5-10% speed for the occasion video compression run.
https://m.vk.com/id207764810 isn't that you?
As far as qualifying how I act, that is a commie tactic. Make stuff up. The only thing I have talked about is Intel's statement and how it is contradicted by the company's practices. Intel has virtual monopoly of the desktop cpu market. Doesn't matter what anyone else in the market does.
Anyway, I don't believe much of what you are saying after going through this with you.
Churka.
Last edited:
What are you doing on this forum then, Mr I do not overclock?
That is a fair question. Mostly it's been to learn about the parts I want because as I've seen, lessons from overclockers translate well to extended stock use reliability well. An overclockers part will outlast ANY mainstream component any day of the week short of enterprise class server hardware, which games rather poorly as you and I both know. Overclocker hardware is the best hardware for a gamer who wants a rig that lasts and lasts and lasts. But even THAT formula comes at the balance of extreme overclockers gear tends to shy away from the direction you want to go with longevity, hence the focus towards the TUF series motherboard instead of ROG, or the non-Extreme Aorus over the Extreme. Cost wasn't an issue, longevity is.
And, to be fair, this is the first machine in 23 years of my 25 years building machines, 2nd machine total ever, that I'm not manually overclocking. My Sandy Bridge machine that this new build just replaced sat at 5Ghz for 5 years without batting an eye. My previous i7 920 ran at 3.8Ghz while my Q6600 G0 stepping ran a cool and smooth 3.6Ghz, a 50% overclock, almost in perfect mirror to my E6600 Conroe which also nailed 3.6Ghz from it's stock 2.4Ghz. I had many colorful and adventurous overclocks before those even. One of them involved cracking open the cartridge of a slot A Athlon and hooking up a Outside Loop Afterburner Gold Finger Device to it to ramp it from 500mhz to 700mhz.
Yes, I overclock. Just not this time, at least so far.
Edit: Minor unintentional lie in above statement. I have done a TINY overclock on CPU and GPU for a 3Dmark Timespy run, but immediately reduced it back to stock afterwards. I just wanted to see what 4.8Ghz CPU and 2050mhz GPU looked like compared to "stock" (aka factory overclocks only).
Last edited:
Thanks for the answer.
D
Deleted member 184142
Guest
I have problems with lots of statements, doesn't mean that lets me ignore them. They can add warranty terms to their own products if they want. Also, they have to mention it, because everyone takes a single statement to cover everything, like people are doing right now. However they reference the spec, so if no limit is mentioned for a give section for a given chip, they means you are more or less open to do what you want. There is however a voltage limit, of 1.52v.
Gotta love arm chair lawyers.
Let me know how your lawsuit goes.
? What are you saying?
D
Deleted member 184142
Guest
You said you have a problem with their statement, which is fine, but their terms are their terms.
And as far as the frequency, they don't list a limit in the spec sheet, they do list a voltage limit however.
Yes, the statement spoke about frequency limits, which is not in the terms.
Insula Gilliganis
[H]ard|Gawd
- Joined
- Feb 6, 2007
- Messages
- 1,468
Thanks Zarathustra[H] for the message as your comments spurred me on to do some more "investigating".. and interesting about the "forum failure" which I never heard of. Not to doubt your memories in any way as info on the Web certainly isn't always accurate, I decided to delve in a bit deeper, most of which, to me, confirms [H] started in 1999 BUT did find a few bits that maybe back up your memories to some degree..
1 - Archive.org's three snapshots from April & May, 1999 are blank and the first snapshot with any archival info is from Oct, 1999, which actually shows [H] as of Feb 29, 2000!! So much for Web info accuracy!!
2 - The [H] news archive on HardOCP.com starts at July 26, 1999.. BUT.. that first posting is about a site redesign.. Kyle wrote " I am in the process of moving all the older content into the newer style page and this will surely take a while" so obviously [H] was up and running before July 26, 1999. Unfortunately this "older content" Kyle mentioned was never posted in the "archives" section or has since disappeared. An archive.org snapshot from Jan, 2010 (the earliest I could find of the link above) doesn't show any additional content before July 26, 1999 posted. BUT.. more on this later!!
3- KB Networks, Inc., which copyrights HardOCP.com, has a start date of 1998. Perhaps Kyle doing the legal & business paperwork necessary before actually launching the site.
4 - I ran Google searches for just the years 1997 and 1998 for the term "hardocp.com".. the 1997 search brought back two results but neither had anything to do with [H].. and the 1998 search gave four results, which only one was significant (unless the link to pics of Asia Carrera can be considered "significant".. and some probably will think so).. Kyle posted a letter sent to him by someone named Lawrence from Intel dated December 10, 1998 about overclocking Intel CPUs and electro migration. This message was sent NOT to [email protected] but to another address, indicating that the HardOCP.com domain wasn't available in Dec, 1998. There are no links or references to why Kyle received this although the message mentions Kyle making a statement.. where this statement was made is unclear. A 1999 only Google search brought up a lot more HardOCP.com hits as it should since we do know the site was operating by then.
So based on what I previously found and this information above, I would conclude that [H] didn't start up until Jan, 1999.. and I was feeling pretty confident in this until.. and here is that BUT from #2.. while doing searches for 1997 and 1998 via https://www.hardocp.com/search/, nothing under 1997 but some hits for 1998.. "Original Voodoo Overclock" dated Jan 01, 1998!! AND "Celeron 333 Review" dated Sept 26, 1998 are two examples, although at the end of the Celeron 333 "review", Kyle again uses a different email address instead of [email protected]. Using [H] search, I was able to find three videocard and five CPU reviews listed before Jan 24, 1999. Unfortunately, there is some uncertainly on whether the dates on these early reviews are accurate as the first two CPU reviews listed as being done by Steve are both dated January 01, 1995. And both reviews are for products that would not be available for many years later (Duron 800 and Pentium III). This "dating problem" does put into question whether any dates before 1999 are accurate. My speculation is that perhaps before [H] was set up, Kyle was posting at another site and brought over some of his reviews/posts when [H] went live.
I still stand that [H] didn't exist in the format we all know today until Jan, 1999 but certainly Kyle had some kind of web presence during 1998. Perhaps that is why you remember "visiting" Kyle in 1998. And of course I could be completely wrong about all of this!! Just ask my wife how often I AM right!!
Things about Hardforum.com are a little different than I first wrote about. There was something called "Public Forum" listed on [H] during most of 2000. The March 2, 2000 Archive.org snapshot has this forum as "under construction" but it appears to be open by May 11, 2000. Clicking the "Public Forum" link took you to https://forum.hardocp.com/, which doesn't work anymore. "Public Forum" was changed to "[H] Forum" sometime between Dec, 2000 and January 2001. When that happened, the link addressed changed from forum.hardocp.com to hardforum.com. I guess Kyle thought that the Forum should be on a seperate site.
I speculate the "Public Forum" was open or at least started accepting members as of March 9, 2000 which you might recall is the join date for [H]er #9 (hemi) as well as all members going into the 500s, then the next join date I can find is March 10, 2000, [H]er #565 (ibenez). If there was some "forum failure" that required signing back up again, it doesn't appear to have affected these early members. So while the actual Hardforum.com didn't exist until Nov, 2000, forums were used in here starting eight months earlier, probably why you "lurking" for awhile before actually joining in Oct, 2000.
FUN FACT - There is a Hardforum.RU in existance!!
FUN FACT 2 - As of today, this site has done 876 videocard reviews and 192 CPU/Processor reviews!!
FUN FACT 3 - Kyle did a "viking funeral" to a deceased motherboard
FUN FACT 4 - The older "currently active" member I can find is SKiZZ, who joined on March 9, 2000 and was active last Thursday.
Hope this was interesting to someone besides myself.. as an undergrad I majored in History so I do enjoy digging around into historical discussions!!
yes, a 1.52 is sounds like they are ok with overclocking, but the statement said you can't run higher frequencies.
D
Deleted member 184142
Guest
It is a statement, it has to be broad, some chips might, some might not, all they did was reference the spec sheet.
However, they can list a limit if they wish.
Hold on then, more wait here. Intel SAID 1.52 vcore is safe or at least specced in? I'm really trying to wrap my head around this as that just seems like a CRAZY amount of vcore for a 14nm chip to survive, let alone run smoothly. That would have blown my Sandy Bridge to pieces! I guess Kaby Lake really is a lot tougher then Sandy Bridge.
DocFaustus
2[H]4U
- Joined
- Sep 22, 2002
- Messages
- 2,830
The first chip I ever overclocked was the famous Intel Celeron 300A. I learned to do that by reading this site. That chip launched in 1998. So this site has to be at least that old.
Yes, I lurked for years before joining as well.
- Joined
- May 18, 1997
- Messages
- 55,601
I was into this late 1998 as I very much remember going to Computex in 1999...maybe '98......hmmmm
Doc is right, I was lurking back then too. I remember the 300A overclocking days fondly, 450mhz was the first 50% CPU OC that was highly likely with most chips. This set the gold standard that I tried to keep reaching for many years and only refound again in the E6600/Q6600 era for awhile.
serpretetsky
2[H]4U
- Joined
- Dec 24, 2008
- Messages
- 2,180
I wouldn't take a datasheet destined for motherboard engineers to be a nod from Intel that you can push your CPU to 1.52 vcore. If you read the notes you will also see that they point out that each processor has an individual VID that should be followed. 1.52V might even be an absolute maximum regardless of clock speed (even if you downclock way below its rated frequency).
If your curious, your sandy bridge also has 1.52V max. Here you go:
https://www-ssl.intel.com/content/d...eets/2nd-gen-core-desktop-vol-1-datasheet.pdf
Scroll down to DC specifications under Electrical specifications.
![Zarathustra[H]](https://cdn.hardforum.com/data/avatars/m/9/9298.jpg?1707758471){kind=avatar}
Zarathustra[H]
Extremely [H]
- Joined
- Oct 29, 2000
- Messages
- 38,835
So, yeah, it turns out HardOCP predates the HardOCP.com domain name by a little bit.
As far as what the original domain name was, no one can quite seem to remember.
probably ocp.<something>.net
I've done some googling and have not been able to turn it up.
rezerekted
2[H]4U
- Joined
- Apr 6, 2015
- Messages
- 3,051
I think you underestimate the hardware knowledge of the average gamer. You talk like most are noobs and never OC. Me thinks that is a wrong assumption.
You assume wrongly. I deal with the "average gamer" daily. I answer their questions, daily. I get looks of shock when I tell them they have to install extra software just to make the LEDs work and they get bent out of shape. No, I won't make the mistake of overestimating them as a group. They are not [H]ard at all. Mention XMP to most of them and they will run away.
rezerekted
2[H]4U
- Joined
- Apr 6, 2015
- Messages
- 3,051
Where do you deal with them daily? I see them in MP games and online forums and most seem to know what is what. It ain't rocket science.
Ummmm, my job? Gaming technical support, contract is NDA. Most gamers I talk to are clueless about what's inside their tower other then "It has a fast Geforce card in it!" and then they get lost when I ask which model. Whats worse is when they tell me about their video card... in a laptop. And then ask me how to upgrade it... no MXM module present or accounted for...
rezerekted
2[H]4U
- Joined
- Apr 6, 2015
- Messages
- 3,051
But those are noobs or kids looking for tech support and not the average gamer. Average gamer are right here on HardOCP and other similar sites. I think most know quite a lot about hardware etc.
Bandalo
2[H]4U
- Joined
- Dec 15, 2010
- Messages
- 2,660
I'd like to think the OC crowd is "above average" gamers.
D
Deleted member 184142
Guest
Those are, indeed, without a doubt the "average" gamer.
People who are on forums like this are very, very few in relation to the number of "gamers". Lots of them OC, sure, but not what WE think of when we say OCing, most mobos today have some form of auto OCing, all you need to know is how to run an .exe file and click a button, it requires no understanding of the system, just as in the thread where these temp problems were posted, someone who actually knew how to OC popped in and started asking questions, 90%+ of the people with the problem reported using the auto OC features and it was adding WAY to much voltage, the guy walked many through setting a manual voltage and they either stopped posting or came back and said "wow, huge change in temps".
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,548
Plenty of people are having trouble even at stock as the thermals jump up and spike and cause the fans to spin up and down. Its a flaw and Intel should stand behind their product rather then blow people off.
As you can see, Rezerekted, I'm not the only one who knows [H] does not represent even a fraction of the PC gamers out there. [H] forum members are dozens of times more knowledgeable then the average gamer. You are far too optimistic, and that kind of optimism sets yourself up for disappointment when you discover the truth of the matter. Gamers are as stupid as any other large group of sheeple. Most of them are lucky to get their computer booted up, let alone actually know what's under the hood.
Wrong. Plenty of people running stock are NOT having these issues to determine this is NOT a chip flaw, it's a motherboard issue, clearly. I, and many others I know, have the 7700K and no problems with temperature spikes. You are off base with your assertion completely.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,548
Ahh the old I don't have a that problem so it doesn't exist motto. I could link forums and I could just stroll over to our own cooling forum and point out people having the problem but your just blind and I cant fix that. Its a chip issue a motherboard can make it worse but the fact is the chip has thermal issues. Since you dont list your speed in your signature I am assuming you dont overclock or your chip cant, otherwise I am not sure why your on here, since we overclock here. People are ticked off for a reason and the lame response by Intel only made it worse.
You point to me being a lone example while you're not EVEN an example at all. You have no clue, you don't have one to speak of! Meanwhile, my tech support job deals with maybe a dozen of these machines a day. Talk out your butt all you want, I'm talking from experience with the chip in question while you quote forum posts...
Edit: In fact... seeing as you DON'T have one, what exactly IS your vested interest in trying so hard to prove this? Seems to me rather questionable how you don't even have this chip to complain about the issue in question, yet you want to insist the whole line is flawed. Seems more like your logic that is flawed here.
Edit: In fact... seeing as you DON'T have one, what exactly IS your vested interest in trying so hard to prove this? Seems to me rather questionable how you don't even have this chip to complain about the issue in question, yet you want to insist the whole line is flawed. Seems more like your logic that is flawed here.
D
Deleted member 184142
Guest
Trouble? You mean seeing a temp number based off a calculation from a sensor that is meant to detect distance to TjMax, that has no change on function, has damaged no chips, and are within spec for the CPU and from the linked forum seem to almost all be the result of auto OCing or users who don't know how to setup a system? Ok....
I think lots of that has to do with how PC gaming used to be, far fewer were into it and building PCs was harder, and OCing required actual knowledge of the system or even hard mods to get an OC. Today HW has really advanced, to the point that it is stupid easy.
However the "idea" of PC gamers is still of the older type.