Delidding a Core i7 6700K

[HdFkK]

After all this thread is was more informative than Linus' video on the topic. 

 

Anyway: Going to get a delidd die Mate and delidd all my CPUs on the very first day after my warranty is over. I so hate the temperature differences on some of the cores of my 5820k. I really can't even look at my core temps for very long because it just freaks me out...

[scottyseng]

5 hours ago, HdFkK said:

After all this thread is was more informative than Linus' video on the topic. 

 

Anyway: Going to get a delidd die Mate and delidd all my CPUs on the very first day after my warranty is over. I so hate the temperature differences on some of the cores of my 5820k. I really can't even look at my core temps for very long because it just freaks me out...

If you have a 5820K, you can't delid it. Haswell-E (And all of the -E CPUs) are soldered.

[Benji_w]

Before delid:

 

Delid:

 

Dunno linus, say thats a pretty good improvement, before result was 1.215v after is well you can see...

[Majestic]

I love how @LinusTech takes such responsibility for his video's, and responds to genuine criticism. He totally doesn't hope people just forget about it in a week while he rakes in the ad-revenue from his ill-researched, misinforming, disingenuous, rushed-out video's to saturate his channel with.

 

You have 2.5million subs, Linus. And thus a large group of people will be exposed to your material (and perpetuate it). The material will be used in recommendations. And who are people going to believe, some forum member or Linus. Debunking your video is nigh-on impossible, because there will always be some asshole who uses a cheap appeal to authority to discredit the counter-argument, rather than argue the point with facts of his own. And since pratically noone will delid their CPU to make a point, only the brave few can debunk your crappy video's. Hardly the critical mass to make a dent.

 

Take some responsibility and respond, atleast. Pull the video, make a follow-up. Anything. It's reached it's viewcount saturation-point anyway. So why would you care. 

 

 

600K views, 96,3% like/dislike ratio. Can you see where i'm going with this, Linus. How much of your audience has literally no knowledge of this and agree with your video's because they're willing to delegate the trust to you. And how much of a critical mass this will reach after it remains up and people perpetuate it.

 

You're causing more problems than solving them with not taking responsibility. Just have a look at the kind of people you're creating based on this dogmatic, doctrine-ridden forum. Everything is echoed around with literally no way of stopping the momentum. Because it only takes one cheap ad-hominem, one appeal to authority or one character assassination by some cognitive-dissonance-suffering capt. Assmad to level the counter-argument you're making. "LOL, so you're saying Linus is making bad video's? LOLOL (+20 thumbs up)". Bam, you're done.

 

 

[Benji_w]

3 hours ago, Majestic said:

Stuff

I agree, also, just going to leave this here.

[techguymaxc]

I posted this in the Youtube comments, but I figure the forum is likely to produce more constructive discussion.

 

Let me start off by saying I am a fan of the channel and have watched hundreds of LTT Youtube videos over the last year or so. Please do not interpret the following as an attack, I only intend to instruct for the purpose of correction.

 

I've been de-lidding CPUs since 2012 when Ivy Bridge was released. I've personally seen temperature decreases in excess of 30C in my own testing with custom water cooling and a lapped IHS & water-block. Many people report temperature decreases under load of 10-15C, with still others reporting anywhere from 20-40C drop.

 

Whenever you construct a testing method to test an hypothesis, you need to ask "why am I doing this/why do people do this?" In this case, the answers are:

1) increased overclocking capabilities

2) reduced temperatures

3) reduced power consumption

 

People de-lid their CPUs because they are thermally-bound when overclocking. In order to push past the limitations of the stock CPU configuration, it is necessary to change one or more parameters. In Ivy Bridge and newer mainstream platform Intel CPUs, the thermal interface material (TIM) between the IHS and the CPU die is the primary limiting factor affecting thermal transfer. Intel themselves acknowledged this with the release of their Devil's Canyon CPUs.

 

With that in mind, the scenario demonstrated in this video fails to account for the primary purpose of de-lidding, i.e. increased overclocking capability. Here then, is everything wrong with your testing method:

1) wrong TIM (you need something better than stock to see real results, i.e. liquid metal compound)

2) stock Intel cooler (way out of its league)

3) no overclocking (you need to raise the TDP of the CPU to realize the maximum benefits of increased thermal transfer)

[ViKT0RY]

Thank you for the Digital Foundry channel discovery. Looks pretty good. I've also delidded my G3258 and don't regret it. I got pissed of this video and the Coollaboratory liquid pro review, they couldn't have done it worse.

My impression is that Intel has told them some directives, as I think the use of the thermal compound is just a way to be better than AMD but have an ace under the sleeve, so they can return to solder and have a performance bump when they want.

[EvilWiffles]

If you spend more than $100 on a CPU cooler, why the hell would you not bother delidding?

Pretty shameful video imo. I remember requesting a video about delidding because I was skeptical about it myself. Luckily I just decided to do it on my new 4790k and really helped a lot. If I watched this without delidding, I probably wouldn't have bothered (that is, if I don't visit Overclock.net often). Now I can enjoy the noiseless H100i, without the fan RPM jumping all over the place and the nice cool temps. I admit, I didn't delid for the overclock headroom but I can certainly get 4.8GHz now if I wanted to without reaching 80c.

[BuckGup]

On 3/4/2016 at 6:05 AM, FilipSebik said:

This is worst delidding i ever saw.

This is best ^ and you cant cut to PCB with this

 

And i guess linus doesn't saw that results from luke's thermal compound application that it doesn't matter unless you put small amount of TIM

Great way to bend the heat spreader!

[Benji_w]

1 hour ago, SamStrecker said:

Great way to bend the heat spreader!

If you're bending an IHS in a vice like that there is something very very wrong with your technique...

[Godlygamer23]

On 3/4/2016 at 2:39 AM, Lolucoca said:

Ummm... Tony Hawk's Pro Skater 2? Wrong video image, I guess @nicklmg

No. It was performed on that surface.

[PlayStation 2]

Can someone tell me how the everloving fuck this went from poor delidding results to AMD?

[Godlygamer23]

On 3/11/2016 at 8:05 PM, techguymaxc said:

1) wrong TIM (you need something better than stock to see real results, i.e. liquid metal compound)

 

2) stock Intel cooler (way out of its league)

3) no overclocking (you need to raise the TDP of the CPU to realize the maximum benefits of increased thermal transfer)

I agree with your second statement 100%. They should've included both a decent AIO or air cooler in their results, and then compared it to just using the stock cooler while delidding.

[Benji_w]

10 minutes ago, Godlygamer23 said:

I agree with your second statement 100%. They should've included both a decent AIO or air cooler in their results, and then compared it to just using the stock cooler while delidding.

At the end of the day it's a shit video full of misinformation, a staple of LTT ethos as of late.

[BuckGup]

18 minutes ago, Benji_w said:

If you're bending an IHS in a vice like that there is something very very wrong with your technique...

Nvm it is against the heat spreader and PCB

[TechnoBeaver]

I had never even heard of de-lidding CPUs, but I even felt something was wrong with that video. Just the way he denounced it felt off.

[Majestic]

It's another episode of Holy $hit! This week: @LinusTech still doesn't take responsibility for his poorly made video!

[PlayStation 2]

39 minutes ago, Majestic said:

It's another episode of Holy $hit! This week: @LinusTech still doesn't take responsibility for his poorly made video!

lol yes

Not sure what's supposed to be cool about [arbitrary expensive shit].

[DildorTheDecent]

That's the worst possible way to delid a processor. 

 

No wonder why Linus doesn't so it. He doesn't even know HOW to do it. 

[PlayStation 2]

2 hours ago, DildorTheDecent said:

That's the worst possible way to delid a processor. 

 

No wonder why Linus doesn't so it. He doesn't even know HOW to do it. 

I mean considering this is the same man who does 52 trillion videos about the Titan X but can't be bothered to do a single 380/390 or interesting GTX 960/970 video...

not surprising in the slightest.

[Majestic]

So, @LinusTech @nickmlg, why no response? I think the criticism is fair...

 

[Opferlamm113]

What I think is most disappointing, a company that sells a high tech high price item cheaps out on a low cost part that would increase the effectiveness and lifespan of its product.

Just take a look where less TDP matters more, like laptops or very small cases where you easily run into the temperature limit.

[Majestic]

33 minutes ago, Opferlamm113 said:

What I think is most disappointing, a company that sells a high tech high price item cheaps out on a low cost part that would increase the effectiveness and lifespan of its product.

Just take a look where less TDP matters more, like laptops or very small cases where you easily run into the temperature limit.

Meh, they function. And delidding actually rarely really increases your maximum overclock. It's more a subective thing, you don't want to see your chip running 80 degrees.

 

But nobody has any data about temperature vs. lifespan.

 

Laptops have such low consumption numbers, that the thermal conductivity of the TIM isn't at all the bottleneck. It's usually the cooling of the laptop and it's thermal conductivity and dissipation.

 

________________________________

 

That said, i'm not defending their TIM. I agree it's shitty.

[HdFkK]

On March 8, 2016 at 7:12 AM, scottyseng said:

If you have a 5820K, you can't delid it. Haswell-E (And all of the -E CPUs) are soldered.

Well, my 6700K isn´t.

[EdgeOfSanity]

7 hours ago, Majestic said:

... And delidding actually rarely really increases your maximum overclock. ...

You couldn't be more wrong.