Kaby Lake crap TIM , WHY ?

[SomeGuy133]

TL: DR he was patently false in his description and back tracked to save face. What he decribed was not the issue as SL correctly explained. It isn't my fault he can;t use basic words to describe something so basic. I am just ensuring bad info from lolfail9001 and Shintai doesn't spread.

He is right.
The major issue is how high the lid is raised off the core, meaning more TIM is needed to fill the gap.
ie the lid is glued in place with a large gap filled with paste.

so silicon lottery is correct that its the application not the glue....he claimed the glue was the issue, which is not accurate.

He said it wasn't TIM (irrlevent) and the glue was going rogue...what the fuck ever that means.

But as he later changed his entire story after silicon lottery stated he is a mook (my word) and that its not the glue but bad application of the IHS due to insufficient pressure during the application of the IHS and crappy TIM. He than completely changed his story...original story below.

He is completely wrong in the statement i quoted. These are his words and they are patently false and inaccurate, which is why i called bullshit the entire time in this whole thread and found silicon lottery accurately and correctly explaining it, which is completely opposite of what he said below.

Don't care what you say you were completely off so just admit it.....your words are below and it is complete bullshit.

.........Also, it was explained thousand times already that it is not TIM at fault, TIM is irrelevant in fact, it is glue going rogue that causes temps.

^^^^^
ummmm no? You said this....did you forget?

Reasonable enough.

Oh, it is. It is not the glue material, and nobody ever claimed that. But glue application is just as freaking essential as any other property of glue, so you may as well consider it part of term "glue" in regards to the complex.

I'll give it that i was not clear enough, but my point was always constant.

I see and hear back tracking.......The TIM is irrelivent!!!! Well yes or no but yes or no but yes or i don't know!?!?!?!

Obviously it is not the glue material, it is the gluing quality hence the gap, for pete's sake, stop being obtuse.

And yes, obviously Intel's TIM is not as good of conductor as LM, nothing usable as TIM is.

He even says this stupid shit. SL flatout stated its due to not putting enough pressure to push the IHS on to the die and not the glue since SL uses basically the same glue with 0 issues.

Please link where it was done for cost cutting. If I remember correctly there was a podcast about it and cost wasn't part of the issue.

Broadwell-E up to 10 cores uses the same die at 246mm2.

http://www.overclockers.com/3step-guide-to-overclock-intel-haswell/

Speaking of temperatures, surely people will be saying “Haswell has the same problem as Ivy Bridge,” referring to the thermal paste issue. Yes, Haswell has thermal paste, but from Very Authoritative People, the TIM is not the problem. As has been posited across the net when people de-lid Ivy Bridge chips with great results, it’s really the black adhesive that’s the culprit. When you cut out that adhesive, it allows the IHS to sit closer to the CPU die, meaning there is less thermal paste through which the heat has to travel, leading to significantly lower temperatures. Intel’s TIM is really quite good, but the manufacturing process leads to that glue being just a little too thick, which is why you see such temperatures.

When you remove the adhesive you also screw the results due to moving the IHS closer to the die. Often plain on top.

on a side note you can delid soldered chips now....not sure why you would. I didnt look too deeply at this yet but thought i would post it.


https://www.overclock3d.net/news/cp...elids_intel_s_core_i7-6950x_broadwell-e_cpu/1

this works on older gens too i guess.

interesting so Intel uses the Indium corp Indium solar preform (assuming he is correct). I was talking to someone on overclock.net about trying Nanofoil on a TEC air/waterblock build. I assume thats not what he is referring to due to this reaching 1500C during application

 

[Dan_D]

you ccan apply 60-80PSI and maybe even more to the die directly in nahalem and IB builds. My 920XM and 3920XM both had 60-80PSI on them from what i could tell. might have been more on certain spots. This was like 4 years ago so it is a little hard to remember.



yes by pissing off customer base to save 25 vents per build instead of just charging an extra fucking dollar for decent TIM?

People wouldnt have an issue with it if they just used a better TIM. I would gladly pay 10 extra dollars to Intel inorder to not have to pay someone 50-75 to do it for me -_-

No matter how you spin it DAN Intel is creating a lot of waste for no fucking reason by cutting an unneeded corner


How much money does Intel pay you to spin that bullshit?Basic logic and economics calls bullshit.

First off, Intel's never given me a dime. I've never received a single penny from a computer hardware manufacturer outside of mail in rebates that I'm rarely diligent enough to send in. Disagreeing with your blown out of proportion complaints is not proof I'm some sort of corporate shill. I disagree with you because I don't think this is as big an issue as you are making it out to be. That's my personal opinion alone.

Secondly, you are making up numbers. Intel doesn't disclose enough about its manufacturing costs to say anything concretely. This is especially true for a product that hasn't officially been released. You are talking out of your ass when you try to give specifics about this or that costing a quarter, or a dollar. Soldering vs. adhesives, crappy TIM vs. a premium TIM may or may not seem like a significant cost difference to you (an assumption on your part) or I because we are buying a single CPU here or there. Most people don't buy more than one every couple of years or so. Even if you bought two or three per year you are no where near a point where that cost adds up. Intel is building and selling tens of thousands of CPUs, perhaps millions all said and done when the dust settles on Kaby Lake and its variants. In addition there are economies of scale to consider. Altering the manufacturing process for one model of CPU has its own costs associated with it.

Could Intel have done things better? Of course! Saving a few cents (assuming that we are actually only talking about pennies here) seems insane to you or I on the surface but when you multiply that by millions of potential CPUs it becomes far more significant. Getting butthurt about this is ridiculous. Do you have any idea how much cost cutting car manufacturers do? Home builders? Virtually everything you buy is built with cost cutting done to maximize profits. Homes use 1/4" sheetrock in most cases when 1/2" is common in commercial buildings. Contractor grade toilet guts fail in three or four years on average, contractor grade fixtures are shit. Commercial buildings use flame retardant sheetrock. Homes do not. Car manufacturers tend to use shitty tires and cheap plastics where they could use better things. Do these things upset you equally?

I'm not sure what you want. Kaby Lake is virtually identical to Skylake on the desktop save for one important detail. Its improved in the most important area for enthusiasts and all you can do is complain? Again, I realize Intel could do better here and that would be great but Intel's going to make whatever decisions it believes will maximize profits and placate shareholders. Its a business, not a charity. Intel probably calculates that for every nerd like you that rages about this, they'll sell a few thousand CPUs to people who don't know any different or simply don't care. This might matter more if Intel had some real competition but as of today you can't buy faster CPUs for desktop computing so there isn't much you can do about it. Intel's provided a minor upgrade in Kaby Lake for overclockers. You can buy it or don't. I don't care and honestly, Intel probably doesn't care if you buy it either. There are plenty of people in the mobile market and server markets who will. We desktop users are a niche market these days. It isn't worth it to Intel to spend money on a solder or a better TIM when the vast majority of these CPUs will never see significant overclocking.

 

[SomeGuy133]

First off, Intel's never given me a dime. I've never received a single penny from a computer hardware manufacturer outside of mail in rebates that I'm rarely diligent enough to send in. Disagreeing with your blown out of proportion complaints is not proof I'm some sort of corporate shill. I disagree with you because I don't think this is as big an issue as you are making it out to be. That's my personal opinion alone.

Secondly, you are making up numbers. Intel doesn't disclose enough about its manufacturing costs to say anything concretely. This is especially true for a product that hasn't officially been released. You are talking out of your ass when you try to give specifics about this or that costing a quarter, or a dollar. Soldering vs. adhesives, crappy TIM vs. a premium TIM may or may not seem like a significant cost difference to you (an assumption on your part) or I because we are buying a single CPU here or there. Most people don't buy more than one every couple of years or so. Even if you bought two or three per year you are no where near a point where that cost adds up. Intel is building and selling tens of thousands of CPUs, perhaps millions all said and done when the dust settles on Kaby Lake and its variants. In addition there are economies of scale to consider. Altering the manufacturing process for one model of CPU has its own costs associated with it.

Could Intel have done things better? Of course! Saving a few cents (assuming that we are actually only talking about pennies here) seems insane to you or I on the surface but when you multiply that by millions of potential CPUs it becomes far more significant. Getting butthurt about this is ridiculous. Do you have any idea how much cost cutting car manufacturers do? Home builders? Virtually everything you buy is built with cost cutting done to maximize profits. Homes use 1/4" sheetrock in most cases when 1/2" is common in commercial buildings. Contractor grade toilet guts fail in three or four years on average, contractor grade fixtures are shit. Commercial buildings use flame retardant sheetrock. Homes do not. Car manufacturers tend to use shitty tires and cheap plastics where they could use better things. Do these things upset you equally?

I'm not sure what you want. Kaby Lake is virtually identical to Skylake on the desktop save for one important detail. Its improved in the most important area for enthusiasts and all you can do is complain? Again, I realize Intel could do better here and that would be great but Intel's going to make whatever decisions it believes will maximize profits and placate shareholders. Its a business, not a charity. Intel probably calculates that for every nerd like you that rages about this, they'll sell a few thousand CPUs to people who don't know any different or simply don't care. This might matter more if Intel had some real competition but as of today you can't buy faster CPUs for desktop computing so there isn't much you can do about it. Intel's provided a minor upgrade in Kaby Lake for overclockers. You can buy it or don't. I don't care and honestly, Intel probably doesn't care if you buy it either. There are plenty of people in the mobile market and server markets who will. We desktop users are a niche market these days. It isn't worth it to Intel to spend money on a solder or a better TIM when the vast majority of these CPUs will never see significant overclocking.

You made my point Dan. I dont need insider knowledge to know the cost behind cheap TIM vs good TIM. You could assume intel gets its TIM for free for all i care and give them the most conservative figure. I could still figure out the maximum at retail Liquid metal costs per die and guess what? Intel could get it cheaper. Still not even 1 dollar per CPU for good TIM if you assume the most expensive figures. That is an irrefutable fact. You can also look at the solder costs and see how much that costs and get an idea of how much more it is though i dont have that knowledge off hand...i am sure its googable or Jim at Indium Corp could tell you.

As silicon lottery already pointed out they have a process issue and use bad materials, which doesn't cost anything to fix really. Hell people do it by hand for 30-50 dollars! It obviously costs a fraction of that.

Auto manufactures using shitty tires? of course it does because i have to replace them with decent one everytime i buy a car which creates economic waste.

Is basic economics too complicated? Some one cuts a corner to save themselves 2 dollars results me spending another 50 dollars to fix their cheapness hurts me and the economy....its called waste....Basic 101 stuff here. So yes when anyone does stupid shit that creates needless waste it irritates me. Like when I buy a home they throw in the shittiest appliances....how about you dont put any in and save me the time and money so i can buy a good product?

 

[KIAman]

I have another take on this.

A chips lid is attached to the substrate with a solidified gel.
The lid is held off the CPU leaving a gap that the TIM fills.

When you delid, the pressure on the lid is moved from the substrate directly to the CPU core in the centre.
This can bend the CPU substrate such that the core warps (my delidded 6700K does this, I got a pm from somebody else with the same problem on his 6700K).
The thinner substrate and weaker sockets have made this easier to occur.
This is a reason Intel use the lid, as stress relief for the CPU core.
It could be why they dont like soldering, there isnt enough stress relief from the core with it being solid.

The delidded CPUs lid now sits directly on the core and gets the full pressure from the heatsink, it has practically no space for the TIM.
This will make it run a stack cooler.
As long as the Silicon doesnt bend which can cause some cores to be hotter than others due to TIM moving and leaving a gap.
I am using a socket 1155 cooler that can apply too much pressure to socket 1151. I found this out the hard way.
I found all cores remain very cool at first but after a while, 2 or 4 (of 8) cores get up to 30C hotter than the others.
The number of hot cores and temp rise varies with how much pressure I put on the heatsink and the exact position of the lid.
I now fear that the socket has bent a little because I always get 2 (of 8) cores that rise a lot higher than the others after a few days when it has been used at temperature.
This is with a tiny amount of pressure from the heatsink.
I can shift which 2 cores by changing the position of the lid on the core.

Thats it.
In order to provide stress relief there has to be a gap.
The TIM fills that gap.
Remove that gap, cooling will improve substantially. But the chip might bend.

I'm going to bed soon so may reply to any lambasting in the morning

When I was delidding my 6700k and stunned by the temperature drop, I had the same realization. I can't fathom just the TIM alone can account for 20-30c drop. However, remove all the glue on the heatspreader forced the heatspreader to make stronger contact against the CPU. Shit, I wish I was aware of this and tried this test. Delid, remove glue, keep original TIM, place heatspeader back on with the glue removed.

 

[Nenu]

Unfortunately air gets into the mix.
No harm trying though.

 

[aaronspink]

You made my point Dan. I dont need insider knowledge to know the cost behind cheap TIM vs good TIM. You could assume intel gets its TIM for free for all i care and give them the most conservative figure. I could still figure out the maximum at retail Liquid metal costs per die and guess what? Intel could get it cheaper. Still not even 1 dollar per CPU for good TIM if you assume the most expensive figures. That is an irrefutable fact. You can also look at the solder costs and see how much that costs and get an idea of how much more it is though i dont have that knowledge off hand...i am sure its googable or Jim at Indium Corp could tell you.

Liquid Metal is not a viable production TIM. PERIOD. FULL STOP. First of all it is electrically conductive. Second, long term reliability is not established. Third, of actual viable TIMs, the ones that Intel actually uses for die to lid interfaces are actually very good.

 

[aaronspink]

When I was delidding my 6700k and stunned by the temperature drop, I had the same realization. I can't fathom just the TIM alone can account for 20-30c drop. However, remove all the glue on the heatspreader forced the heatspreader to make stronger contact against the CPU. Shit, I wish I was aware of this and tried this test. Delid, remove glue, keep original TIM, place heatspeader back on with the glue removed.

It is important to point out that the Lid is a stress relief mechanism. The tolerances used are designed to increase overall product reliability in the consumer space. It should therefore not be shocking that Intel aren't using the minimal tolerances possible. They are using the tolerances that results in the highest reliability across the product line with minimal impact on performance.

 

[DWolvin]

Is basic economics too complicated? Some one cuts a corner to save themselves 2 dollars results me spending another 50 dollars to fix their cheapness hurts me and the economy....its called waste....Basic 101 stuff here. So yes when anyone does stupid shit that creates needless waste it irritates me. Like when I buy a home they throw in the shittiest appliances....how about you dont put any in and save me the time and money so i can buy a good product?

I think I understand what you want to say, but the company saving $2 gives them an extra $2. They don't care at all how much you may or may not spend to fix it later. And as a side not, you spending money to fix it hurts your personal funds, but is good for the economy.

 

[lolfail9001]

TL: DR he was patently false in his description and back tracked to save face. What he decribed was not the issue as SL correctly explained. It isn't my fault he can;t use basic words to describe something so basic. I am just ensuring bad info from lolfail9001 and Shintai doesn't spread.

See, you have just illustrated my entire fucking point.

glue going rogue

Tell me with straight face this statement could ever be about anything other than application process. I mean sure, if you believe i believe glue has intellect, then you have a point.

on a side note you can delid soldered chips now....not sure why you would. I didnt look too deeply at this yet but thought i would post it.

It drops temperatures by 10 degrees on Haswell-E, in case you did not know. Just industrial scale application process things, applies to both gluing AND soldering.

He even says this stupid shit. SL flatout stated its due to not putting enough pressure to push the IHS on to the die and not the glue since SL uses basically the same glue with 0 issues.

gluing quality

Telling you, stop being obtuse.

 

[SomeGuy133]

Liquid Metal is not a viable production TIM. PERIOD. FULL STOP. First of all it is electrically conductive. Second, long term reliability is not established. Third, of actual viable TIMs, the ones that Intel actually uses for die to lid interfaces are actually very good.

proof? ah thats right....none

I think I understand what you want to say, but the company saving $2 gives them an extra $2. They don't care at all how much you may or may not spend to fix it later. And as a side not, you spending money to fix it hurts your personal funds, but is good for the economy.

waste is never good....goodness gracious reread what you just said. That is like saying the government running a useless program helps the econemy by providing jobs lol. If you believe that........yea i am just stopping there

See, you have just illustrated my entire fucking point.

Tell me with straight face this statement could ever be about anything other than application process. I mean sure, if you believe i believe glue has intellect, then you have a point.

It drops temperatures by 10 degrees on Haswell-E, in case you did not know. Just industrial scale application process things, applies to both gluing AND soldering.


Telling you, stop being obtuse.

illustrate what?? you dont know wtf your talking about?

Sorry i dont speak crypic nonsense so glue going rogue...i have no clue*. It bled into the TIM, it acts as a heat trap...i dont know...i dont read brains that can't explain basic shit.

Learn how to form coherent statements with actual logical facts. I am not being obtuse i quoted your incomprehensible garble that was patently wrong. Stop being in denial and face reality.

Also go ahead and quote your statements of back tracking as your original point...lol You are shameless on trying to save face.

I made my point you were wrong and proved your back tracking. I am done.

 

[aaronspink]

proof? ah thats right....none

Sure lets put a conductive liquid with little data right next to sensitive electronics that need a documented 10 year reliability, I'm sure nothing will go wrong... Electrically conductive TIMs aren't viable for production and never have been. No liquid metal TIM provides any warranty and places like SL basically don't provide one either. That's because they are both inherently risky products with unknown long term reliability. There isn't a single manufacturer that uses electrically conductive TIMs anywhere near electrical components.

If your argument that Intel is cheap because they don't use a TIM with known electrical conductivity issues, then your argument is completely invalid.

 

[SomeGuy133]

Sure lets put a conductive liquid with little data right next to sensitive electronics that need a documented 10 year reliability, I'm sure nothing will go wrong... Electrically conductive TIMs aren't viable for production and never have been. No liquid metal TIM provides any warranty and places like SL basically don't provide one either. That's because they are both inherently risky products with unknown long term reliability. There isn't a single manufacturer that uses electrically conductive TIMs anywhere near electrical components.

If your argument that Intel is cheap because they don't use a TIM with known electrical conductivity issues, then your argument is completely invalid.

you claimed that Intels TIM was good so whats the specs? they could easily use other quality TIMS like ICD or anything else thats not conductive. Also they use solder for the others which is conductive......

Sounds like you are making this up as you go.

 

[Nenu]

you claimed that Intels TIM was good so whats the specs? they could easily use other quality TIMS like ICD or anything else thats not conductive. Also they use solder for the others which is conductive......

Sounds like you are making this up as you go.

You are claiming Intel TIM is bad, what are the specs?

Note that solder does not move once set.
It cannot cause later issues.

 

[thesmokingman]

You are claiming Intel TIM is bad, what are the specs?

Note that solder does not move once set.
It cannot cause later issues.

If it were great we wouldn't have gains like this and its not an isolated case.

Before

After

And this is Prime 28.10 running Blend with memory at 3733mhz.

 

[aaronspink]

you claimed that Intels TIM was good so whats the specs? they could easily use other quality TIMS like ICD or anything else thats not conductive. Also they use solder for the others which is conductive......

Sounds like you are making this up as you go.

Intel sources their thermal compounds from Dow Corning last I heard, the same place that pretty much every PC TIM "maker" sources their thermal compounds from. ICD doesn't have better performance than many other greases out there. TC-5622 provides the same or better performance depending on the metric, is extremely well characterized and tested, etc. The idea that Intel is using some random shitty lowest cost basic thermal grease is complete fantasy.

They use solder in lower heat flux and lower heat cycle applications on higher cost packages. That has absolutely nothing to do with them using a grease based TIM in high heat flux/cycle applications. And while solder can be conductive, it is not a liquid and is not effected by things such as pump out.

 

[noko]

For argument, let say Intel puts this mystery best TIM in their Kaby Lake processor - it lowers cpu temperatures 10c with the best cooler out there - How will that really affect the overall OC? By how much? 100mhz - 200mhz? if that much.

 

[Nenu]

If it were great we wouldn't have gains like this and its not an isolated case.

Delidding works, no dispute.
Understanding why is fundamental for a discussion on topic.

 

[Aluminum]

Intel sources their thermal compounds from Dow Corning last I heard, the same place that pretty much every PC TIM "maker" sources their thermal compounds from. ICD doesn't have better performance than many other greases out there. TC-5622 provides the same or better performance depending on the metric, is extremely well characterized and tested, etc. The idea that Intel is using some random shitty lowest cost basic thermal grease is complete fantasy.

They use solder in lower heat flux and lower heat cycle applications on higher cost packages. That has absolutely nothing to do with them using a grease based TIM in high heat flux/cycle applications. And while solder can be conductive, it is not a liquid and is not effected by things such as pump out.

Nobody intelligent thought they were buying barrels of noname TIM to save $0.002 per cpu.

But something is obviously off in their entire process, if anyone has doubts about that by now then I question their motives and/or intelligence.

Maybe it is assembly/tolerances/??? but there are plenty of rational people using proper coolers with correct specifications and mounting methods that have shown large improvements from correcting whatever flaws are in intel's 'consumer' process of attaching the heat spreader.

It is a fair question to ask why is intel not correcting this after several generations of cpus.

Yes most are still "within spec" but with marginal coolers (also "within spec" of their own requirements) some of these consumer assembled cpus are definitely going to throttle in some scenarios. I don't believe their army of mechanical/etc engineers would let this go on as long as it has without a significant cost savings/issue of some kind.
I also find it odd that some 'premium' (K series) and 'enterprise' (E3 xeons) also suffer from this, as both of these product segments are a lot more likely to encounter throttling scenarios.

 

[Shintai]

The question is if there is an issue at all. Would a smaller gap be too little for all the QA? How are other makers using TIM doing gap wise? Just because it works for "me" doesn't mean it works for everyone else.

 

[SomeGuy133]

The question is if there is an issue at all. Would a smaller gap be too little for all the QA? How are other makers using TIM doing gap wise? Just because it works for "me" doesn't mean it works for everyone else.

the fact SL shows results and has turned a business out of intels biff shows theirs a problem that needs fixing.

 

[thesmokingman]

the fact SL shows results and has turned a business out of intels biff shows theirs a problem that needs fixing.

I'm convinced you will have ppl debating the point even in the face of clear and convincing evidence to the contrary.

 

[Shintai]

the fact SL shows results and has turned a business out of intels biff shows theirs a problem that needs fixing.

You mean after it became extremely easy to delid with Ivy Bridge? What a surprise, who would ever have imagined that. (sarcasm)

And before you mention the "good old SB days". Then remember heat density have almost doubled since for the die.

 

[aaronspink]

But something is obviously off in their entire process, if anyone has doubts about that by now then I question their motives and/or intelligence.

Maybe it is assembly/tolerances/??? but there are plenty of rational people using proper coolers with correct specifications and mounting methods that have shown large improvements from correcting whatever flaws are in intel's 'consumer' process of attaching the heat spreader.

It is a fair question to ask why is intel not correcting this after several generations of cpus.

Because it is likely nothing is actually off in their entire process. It is perfectly reasonable to assume that the tolerances they are using are for entirely legitimate reasons. They are literally dealing with hundreds of millions of parts per year and likely have both significantly more data and more importantly significantly more accurate data than anything the community can come up with a couple thousand units. Hell, the data they are concerned about may not even ever be generated by the enthusiast community due to a reliance on aging for which they can actually test and have data.

If Intel is not "correcting" it after several generations, there are likely very good reasons besides saving a couple pennies.

 

[aaronspink]

the fact SL shows results and has turned a business out of intels biff shows theirs a problem that needs fixing.

No it doesn't. That's like saying that because there are aftermarket ECU mods that result in more performance that shows there are problems in automaker engines. All it really says is that SL has a different optimization and reliability point than Intel. And that's perfectly fine. Across many industries there have always been aftermarket mods that result in more performance as a tradeoff for some other attribute. For SL, they basically reduce the die to lid tolerance resulting in more performance. We don't have anywhere near enough data to say what other tradeoffs that entails.

 

[noko]

How much more performance are we looking at?

If Kaby does 5ghz as is but then can do 5.2ghz delided - That is a whole whopping 4% clock speed increase with probably less then 4% performance increase. That to me would be pointless but each their own.

 

[SomeGuy133]

How much more performance are we looking at?

If Kaby does 5ghz as is but then can do 5.2ghz delided - That is a whole whopping 4% clock speed increase with probably less then 4% performance increase. That to me would be pointless but each their own.

depends the person for me i will take every 100Mhz i can get.

 

[opfreak]

depends the person for me i will take every 100Mhz i can get.

so you expect intel to spend millions so that you and a handful of others that spend 300 bucks on a cpu can get what will amount to almost no performance increase.

 

[SomeGuy133]

so you expect intel to spend millions so that you and a handful of others that spend 300 bucks on a cpu can get what will amount to almost no performance increase.

wtf are you talking about? I just made the comment that i am someone who upgrades for that extra 200MHz....

if i had the money at the moment i would swap my 4.8 GHz skylake for a binned Kaby and do phase chamge for 5.4-5.5GHz but i dont have the funds at the moment.

99% of what i do is single thread.

 

[opfreak]

wtf are you talking about? I just made the comment that i am someone who upgrades for that extra 200MHz....

if i had the money at the moment i would swap my 4.8 GHz skylake for a binned Kaby and do phase chamge for 5.4-5.5GHz but i dont have the funds at the moment.

99% of what i do is single thread.

no you are complaining that intel went cheap.

So again, to you, and all the other whiners. Do you expect intel to spend tons of money to get minimal performance benefit, for a handful of people?

 

[w.rong]

Ultimately too little people care in the grand scheme of things, we're talking a subset of people buying higher-end consumer CPUs at retail, which itself doesn't match the volume of OEM sales.
Since most people really won't notice, even the extra 50cents turns out to be a big dent on intels finances, besides it forces "enthusiasts" to buy the soldered extremes.

 

[Activate: AMD]

so you expect intel to spend millions so that you and a handful of others that spend 300 bucks on a cpu can get what will amount to almost no performance increase.

Of course the irony is that many of these people would probably be foaming if Intel changed their process to change the tolerances or whatever and was forced to charge $50 more as a result. Taking 30+ minutes to delid and reapply TIM to "fix" intel's "problem" and consistently ending up with better temps doesn't imply there was a problem at all. It just proves intel has a process with consistent results. That these consistent results aren't some small niche's idea of optimal doesn't mean there is an issue with the process. If hand applying TIM was the only way to achieve these results, do people really expect them to do it? give me a break.

 

[Peter2k]

no you are complaining that intel went cheap.

So again, to you, and all the other whiners. Do you expect intel to spend tons of money to get minimal performance benefit, for a handful of people?

Well to be fair
Intel used to solder, and they still do some parts
So it can't be huge margins

I understand why they would want to save
the TIM works well enough under normal circumstances, like boosting

Still
K CPU's are for OC'ing

I remember people still delidding soldered CPU's as well
But the difference was something like 5° Celsius under load

I've seen reviews where delidding and replacing the TIM dropped Temps by over 20°

Even if you don't OC
15° could be the difference between a silent system and a loud one

--------------------

Mmm
You're saying it's too much of a cost
Well I'll meet you halfway there
Why not sell me a 7700K-X
With no heatspreader attached, but comes with one
For self assembly

If I want to use one

We're talking about overclockers and OC cpu's after all

 

[opfreak]

Well to be fair
Intel used to solder, and they still do some parts
So it can't be huge margins

I understand why they would want to save
the TIM works well enough under normal circumstances, like boosting

Still
K CPU's are for OC'ing

I remember people still delidding soldered CPU's as well
But the difference was something like 5° Celsius under load

I've seen reviews where delidding and replacing the TIM dropped Temps by over 20°

Even if you don't OC
15° could be the difference between a silent system and a loud one

--------------------

Mmm
You're saying it's too much of a cost
Well I'll meet you halfway there
Why not sell me a 7700K-X
With no heatspreader attached, but comes with one
For self assembly

If I want to use one

We're talking about overclockers and OC cpu's after all

Having both a K, and K-X, means now you have 2 products. 2 sku's, 2 boxes, 2 things to sell, 2 things to support, etc etc.

Its just not worth it to intel. The OC market is just big enough to support, but not much more then that.

 

[GregTombstone]

Remember when you could get a 40% + overclock, Bartons and Core 2 6600s, that is what its about , not a piddling 10% on a 500 euro cpu

 

[Bandalo]

Remember when you could get a 40% + overclock, Bartons and Core 2 6600s, that is what its about , not a piddling 10% on a 500 euro cpu

Manufacturers realized it was stupid to leave that much extra overhead on the table. They got better at binning their parts, so there's less room for OC.

 

[Activate: AMD]

Remember when you could get a 40% + overclock, Bartons and Core 2 6600s, that is what its about , not a piddling 10% on a 500 euro cpu

Pretty cynical outlook that doesn't really add up IMO. We're talking about 10+ years of technological advancement, billions more transistors, 2x or more clock speeds and (critically) much much higher heat density. Not to mention designs optimized so much more for TDP than clock speed since these cores are being used in everything from tablets to data centers.

And the Q6600 launched for more than a k-series i5 costs today... it was just that the top end CPU's were just slightly faster versions of the same shit for 3x the cost. Now the CPU that costs $1000 actually has more cores and differentiating features (though its tough to say its a bargain, you at least could argue where the money goes)

Manufacturers realized it was stupid to leave that much extra overhead on the table. They got better at binning their parts, so there's less room for OC.

This too

 

[opfreak]

Pretty cynical outlook that doesn't really add up IMO. We're talking about 10+ years of technological advancement, billions more transistors, 2x or more clock speeds and (critically) much much higher heat density. Not to mention designs optimized so much more for TDP than clock speed since these cores are being used in everything from tablets to data centers.

And the Q6600 launched for more than a k-series i5 costs today... it was just that the top end CPU's were just slightly faster versions of the same shit for 3x the cost. Now the CPU that costs $1000 actually has more cores and differentiating features (though its tough to say its a bargain, you at least could argue where the money goes)



This too

So much the price. q6600 launched at $851, before starting to drop.

 

[Modboxx]

Anyone know the name of the liquid metal product? or link? Also what kind of glue does he use to relid?

 

[Peter2k]

Anyone know the name of the liquid metal product? or link? Also what kind of glue does he use to relid?

Coollaboratory
Liquid Pro/Ultra Thermal Compound

https://www.amazon.com/gp/aw/d/B001...quid+ultra&dpPl=1&dpID=31voj9YA04L&ref=plSrch

they sell the delidding tool in a set
the tool
liquid ultra
and some silicon glue

or used to
from caseking.de
caseking I think is partners with overclockers.co.uk

the silicon glue just reads
"heat resistant silicone"

they don't sell the tool any more

but it's side for glueing it back on doesn't fit kaby lake
I know because I tried

so maybe they just got rid of this variant and make a new one for kaby

I used some heat resistant silicon
but made it a lot thinner then he did
a lot


word of advise
if you actually use some liquid metal, then try it out first on some other metallic "test subject"

it handles very different then anything else

 

[FrgMstr]

Be careful when delidding.