Intel's TIM continues to create heat bottleneck, many years after Ivy

[crashtech]

Intel should just sell delidded CPUs already.

It's a very simple case of liability exposure. Even as a small businessman in flyover country I know better, you expect Intel to be so stupid? The standard of the industry for desktop and server is to have an IHS now, it would be really difficult to regress without being liable for every damaged die.

 

[superstition]

Well, the general consensus in that the new polymer did allow DC to run a few degrees cooler, but obviously not enough to make you happy.

Refer to the article I posted that says the "new TIM" doesn't perform any better than the "old TIM". I put it in quotes because we have to take Intel's word for it being changed in the first place. The result that article discusses suggests it wasn't changed.

As far as the "E" chips, they are perfectly available to buy, dont know why you would say they arent.

Because they aren't. There are no Broadwell or Skylake E chips on the market.

And, those chips are not necessary at all for enthusiast gaming. The extra cores are only a drawback.

Admittedly it would be nice if K chips had solder instead of TIM.

It's not about it being nice. It's about an expensive product, like an i7, being good enough. People who want low thermal transfer performance have Pentiums, i3s, and i5s to purchase. Parts that are expensive and designed for performance like a quad i7 should not be bottlenecked 16-20+ C by polymer TIM.

It was pointed out that segmentation is important for Intel's pricing. Having soldered quads improves the difference between an i5 and an i7.

From a production standpoint though, I am not sure how practical/expensive it is to do this.

Trivial.

If it were easy and relatively inexpensive, I would think intel would do it and sell the chips at a higher price.

You're neglecting the clear incentive Intel has to create the delidding practice.

Or maybe there just arent that many users that would be interested.

Refer to the poll results.

 

[VirtualLarry]

It's a very simple case of liability exposure. Even as a small businessman in flyover country I know better, you expect Intel to be so stupid? The standard of the industry for desktop and server is to have an IHS now, it would be really difficult to regress without being liable for every damaged die.

Why would that be so? I believe that all Intel CPUs, whether boxed or tray, are intended to be "installed by professionals".

 

[VirtualLarry]

You're neglecting the clear incentive Intel has to create the delidding practice.

That sounds a bit ... conspiratorial to me. I really doubt Intel has a "cash cow" of enthusiasts failing to properly de-lid their CPUs and ordering new ones.

If they were really counting on a certain percentage of double-sales to people that are destroying their chips, wouldn't they want to sell already de-lidded CPUs, such that more people might be likely to crack their dies installing them?

 

[crashtech]

Why would that be so? I believe that all Intel CPUs, whether boxed or tray, are intended to be "installed by professionals".

It's obviously not what happens in practice. It would be a PR disaster either way.

 

[frozentundra123456]

Refer to the article I posted that says the "new TIM" doesn't perform any better than the "old TIM". I put it in quotes because we have to take Intel's word for it being changed in the first place. The result that article discusses suggests it wasn't changed.

Because they aren't. There are no Broadwell or Skylake E chips on the market.

And, those chips are not necessary at all for enthusiast gaming. The extra cores are only a drawback.

It's not about it being nice. It's about an expensive product, like an i7, being good enough. People who want low thermal transfer performance have Pentiums, i3s, and i5s to purchase. Parts that are expensive and designed for performance like a quad i7 should not be bottlenecked 16-20+ C by polymer TIM.

It was pointed out that segmentation is important for Intel's pricing. Having soldered quads improves the difference between an i5 and an i7.

Trivial.

You're neglecting the clear incentive Intel has to create the delidding practice.

Refer to the poll results.

You really are determined to make a big deal about this arent you? You basically distorted every comment I made. And hardly anything you said makes sense. First, you can believe intel lied about changing the tim if you want. Seems unlikely to me, because they would face legal action, but hey, if it fits your agenda, knock yourself out. Secondly, HEDT chips are available. You did not specify what generation they had to be. And with modern games, and DX12 coming, more cores are not a disadvantage, in fact they are an advantage, and most likely will become even more so.

Also can you document how much profit intel makes from these delidded chips? I hardly think the results of a poll of less than a hundred people proves anything relative to the millions of chips intel sells per year. And even then, your poll only asks whether posters would prefer solder, not whether they have purchased chips to replace chips damaged by delidding.And like I said, if intel chips are so "bottlenecked" by the TIM that they are not "good enough" for you, feel free to buy Soldered HEDT chips or soldered AMD chips.

 

[Excessi0n]

It's a very simple case of liability exposure. Even as a small businessman in flyover country I know better, you expect Intel to be so stupid? The standard of the industry for desktop and server is to have an IHS now, it would be really difficult to regress without being liable for every damaged die.

Delidded isn't the same as bare-die. They could still throw the IHS in the box and say that you won't be covered by warranty if you don't use it. It's not like the temperature difference between bare-die and IHS with liquid-metal TIM is more than a couple of degrees.

While we're daydreaming about things that will never happen, wouldn't it be cool if Intel printed the chip's information around the perimeter of the IHS so that lapping wouldn't void your warranty?

 

[Dufus]

Intel should just sell delidded CPUs already.

They do, trouble is they are BGA. Be careful what you wish for. 😉


Hard to find example from https://www.youtube.com/watch?v=8GFPiCVM59Q

i5-5675R Broadwell Desktop CPU

 

[Phynaz]

I'd like to know what CPU the op is currently running.

 

[rgallant]

I'd really like to know how you've broken the laws of physics by having your chips run cooler than ambient.

3xMCR320[basement] in sig = 9 x 120 mm of rad + 9 x 120 mm 1600 rpm fans

stone foundation [ 1855 house] = crawl space is cooler than my office
normal ground temps are 55f , think root cellar
also the 2 pumps are down there
plus a 2 nd 500 watt psu [turned on by a 12v relay from the main psu]
that powers all the fans ,pumps and two fan controllers in the pc .[takes load off the main psu so it's fan does not turn on/ramp up]

and no sound as a plus , think small dell optipex
less heat dump in my 8x10' room
all case fans 500-600 rpm
all ssd's

4 slow moving 120mm gt's in the side ,1 200mm in the front
1 140 mm in the back
1 140 mm under the main psu so it's fan only comes on during gaming.

 

[ehume]

which is to say that his true ambient -- the rads' ambient -- is not the ambient of his room.

 

[superstition]

That sounds a bit ... conspiratorial to me. I really doubt Intel has a "cash cow" of enthusiasts failing to properly de-lid their CPUs and ordering new ones.

A corporation's job is to conspire against consumers, giving them a minimum of product for a maximum of profit. Corporations are not non-profits; they aren't charities. They are supposed to enrich shareholders.

The only thing they have to do is provide the minimum product they can get consumers to pay for. Generally as one moves from necessity toward luxury the profit margin increases but the same rule holds true. They don't want to alienate their customers so that they will not buy or will buy something which has lower margin. So, since companies want to give people as little product as possible for the money they have the incentive to try to entice people into the highest margin products. There aren't many things with higher margins than a product that people are going to want to break to improve its performance. Not only does it get rid of the warranty it provides the possibility of total breakage and repurchase.

It's not realistic to see a corporation as a benevolent entity when it comes to consumer desire, only when it comes to shareholder desire (and the desire of the CEO class to become rich with things like golden parachutes).

If they were really counting on a certain percentage of double-sales to people that are destroying their chips, wouldn't they want to sell already de-lidded CPUs, such that more people might be likely to crack their dies installing them?

It's a lot easier to provide no warranty for delidding than it is to try to provide a warranty for a product with no installed spreader. Also, delidding is more dangerous than putting on a spreader after using more effective TIM, especially given the thinned substrate.

 

[crashtech]

Delidded isn't the same as bare-die. They could still throw the IHS in the box and say that you won't be covered by warranty if you don't use it. It's not like the temperature difference between bare-die and IHS with liquid-metal TIM is more than a couple of degrees.

While we're daydreaming about things that will never happen, wouldn't it be cool if Intel printed the chip's information around the perimeter of the IHS so that lapping wouldn't void your warranty?

Actually, delid-relid is more similar to bare die than you might think. The sealant around the perimeter of the IHS is key to ensuring the die is not put under excessive or uneven pressure by the HSF. In other words, the sealant is load-bearing. Selling a CPU without the sealant removes one of the big reasons for the IHS to be there in the first place.

 

[DrMrLordX]

Selling a CPU without the sealant removes one of the big reasons for the IHS to be there in the first place.

That's why Intel and/or AMD should leave a half-thick layer of sealant to serve as a guide for the IHS if anyone wants to use the one included loose in the package (assuming they ever sell a CPU with a loose IHS). It not only serves as a guide for where to put on the IHS, but it prevents it from rolling around on the PCB during mounting of the HSF, and it can form a weak bond due to the pressure of the mount to make handling of the chip easier afterwards. It's quite easy to remove with a simple twist, but it does stick on there.

 

[superstition]

It seems a lot more practical to ask for:

a) One i7 quad that has solder
b) That all i7 quads have solder
c) That all i5 or better chips have solder

versus asking for parts without installed spreaders

 

[superstition]

Just happened upon this when I was trying to find Skylake's transistor count:

To paraphrase Splave again, he comments that the thermal paste (TIM) on his Skylake is certainly worse than that of Devil's Canyon. If the extra mass on the IHS is coming from a taller heatspreader (by virtue of the smaller package substrate), then more TIM is needed otherwise there will be substantial air bubbling of the TIM between the CPU and the heatspreader. By replacing his own thermal paste and resecuring the heatspreader, he saw an 18°C drop in temperatures at his highest air overclock with the old paste (5.1 GHz at 1.48 volts) - from 96ºC that overheated to 78ºC on the warmest core. An 18°C drop is immense. Under those conditions, and based on rough testing not published in our Skylake review, it could equal another 100-400 MHz depending on the quality of the processor. PCWatch confirms that switching out the paste with CoolLaboratory’s Liquid Pro (a liquid metal adhesion interface) reduced temperatures at 4.6 GHz from 88ºC to 68ºC

It is worth noting that previously on certain platforms Intel had been providing a mixed metal interface (generalized as a soldered interface) between the silicon and the heatspreader, which is the best but most expensive option. If the cost of the interface is reduced by 0.1 cents, then that's a significant saving on millions of processors.

When customers have no choice.

http://www.anandtech.com/show/9505/skylake-cpu-package-analysis

 

[LTC8K6]

I would think Intel might have used the better interface method/material on the unlocked multiplier chips, which are probably relatively low volume.

I think there must be more to it than simply saving money.

 

[Burpo]

"Delidding an Intel Core i7 6700K - Is it worth it!?"

https://www.youtube.com/watch?v=LvS_qgn7OSs

 

[superstition]

I would think Intel might have used the better interface method/material on the unlocked multiplier chips, which are probably relatively low volume.

I think there must be more to it than simply saving money.

Why? Consumers are completely at their mercy.

 

[Burpo]

https://www.youtube.com/watch?v=LvS_qgn7OSs

As seen in the above video as he's delidding, intel used sufficient amount TIM, and was making good transfer to heat spreader. There's nothing sinister about what they've done. They're producing quality products, that feeds the masses.

Point fingers, make Polls, get worked up about it, whatever.. I didn't vote in the poll, because IMO it's a non-issue.

 

[Deders]

Not convinced by that video. His results differ greatly from most others.

It has been agreed that it's not the Tim itself that's the problem, but the gap the Tim needs to fill. This is why removing the adhesive is important as it decreases that gap.

 

[Burpo]

It has been agreed that it's not the Tim itself that's the problem

Where's the problem? I don't agree there is one..
Do we see Skylakes that are overheating? Nope. Many running at 4.7GHz without doing anything to the chip.

"Intel's TIM continues to create heat bottleneck, many years after Ivy"

Poppycock!

 

[frozentundra123456]

https://www.youtube.com/watch?v=LvS_qgn7OSs


Point fingers, make Polls, get worked up about it, whatever.. I didn't vote in the poll, because IMO it's a non-issue.

Well said.

 

[Deders]

Judging from results I see in other forums, and Idontcare's thread in this very section, there can be huge gains to be had by delidding, especially when overclocking.

My current 6700k can't dissapate heat effectively enough for my heatsink to do its job properly. The gap is effectively a bottleneck for heat, stopping too much of it from even getting to the heatsink.

If you had a gap between the IHS of a soldered chip and the heatsink you would see similar results.

 

[superstition]

https://www.youtube.com/watch?v=LvS_qgn7OSs

As seen in the above video as he's delidding, intel used sufficient amount TIM, and was making good transfer to heat spreader. There's nothing sinister about what they've done. They're producing quality products, that feeds the masses.

Point fingers, make Polls, get worked up about it, whatever.. I didn't vote in the poll, because IMO it's a non-issue.

Where's the problem? I don't agree there is one..
Do we see Skylakes that are overheating? Nope. Many running at 4.7GHz without doing anything to the chip.

"Intel's TIM continues to create heat bottleneck, many years after Ivy"

Poppycock!

That's your opinion. A lot of others think it is an issue.

If you had a gap between the IHS of a soldered chip and the heatsink you would see similar results.

I don't know of any hard evidence that shows soldered chips, in general, perform equally as bad as Intel's TIM.