Reviews Intel Core i7-8700K (Coffee Lake) And Z370 Chipset Review

I like how the power/temperature tables have turned. Now it's Intel who are chewing through the watts and supplying the toaster-chipset!

 

The problem is that we can pretty much sum up Intel mainstream CPUs of the last 6 years with below:
Everyone's darling chip I5-2500K launched back in 2011.
Then a whole bunch of teeny tiny upgrades and price hikes not worth upgrading for.
Now they suddenly pull the i5-8600K out of the bag.
Even if Ryzen didn't exist it would make no sense that it took them this long, because in the same time the HEDT chips progressed from the i7-990X to the i9-7980Xe.

 

"Let's focus on IGP improvements, because everyone wants that. We'll spend silicon budget making the IGP bigger than the CPU cores"

Also 8700 non-K > 8600K tbh. 6/12 + more cache + high clocks + 65W for 50USD. You'd probably save that much on a H370 than Z370

 

but it's fairly clear that this isn't just a software trick to boost sales; there are real physical reasons why you won't be able to use a Coffee Lake CPU in a Z370 motherboard.

^^Looks like a small Typo here, don't you mean you can't use a Coffee Lake CPU in a Z270 board.

 

There are real physical reasons why CL will not work in Z270 according to Intel.

Just like how they can't use solder since Sandy, yet all of their Xeon range has been soldered and Derbauer also managed to solder a Skylake. He did fail a couple of times but hey, he's not Intel. Funny how the entire Broadwell E and Haswell E range were soldered, though.

So yes, if you believe Intel there is no way CL could have worked on Z270. But I am not buying it, much like I don't buy much else of their nickel and dime crap. If AMD can get Bristol Ridge running on AM4 then Intel could have let two more cores and threads work on z270. Bristol Ridge is not Ryzen it's a completely different core tech, yet AMD somehow managed to get it working on AM4 just fine.

 

Not the E3 line that use the same dies as the Socket H chips, those use identical assembly and TIM. Xeon-W (the Socket R4 Xeon line) also uses TIM.

He did not succeed at all, as he himself said was likely. Ivy Bridge and later lack the additional metallisation layers on the die and underside of the IHS required for a: the solder to properly wet it, and b: not destroy the die by infiltrating it with hot Indium and catastrophically redoping the transistor layer (which is at the top of the die due to the flip-chip production process)

 

Yeah, but, edzieba, I think you'll find that that is, like, totally Intel's own fault for being so bad and stuff.

Or something.

 

Already had this argument elsewhere. Not going there again.

 

Sorry for a quick bump, but where would a 4790k fit in this review?

 

The non-solder thing is the highest order of marketing spin bollocks.
Micro-fractures may well be an issue, but of all the years Intel has soldered consumer CPUs (which have a limited warranty anyway), and the fact they still solder Xeons and AMD solders its chips means its still feasible, yet, they don't want to do the few extra steps of process for cost and yield. Not even for premium K or X chips. Bean counters making decisions.

 

Coming back to this now that I have a little energy (because I had been arguing about it for three days)

https://overclocking.guide/the-truth-about-cpu-soldering/

And most notably the below.

"I spent months and some $$$ on figuring out how to solder a recent Intel Skylake CPU. After reading all the above you can imagine that it’s by far not as simple as it sounds like. I’m not going to publish all info on Skylake soldering for now because I’m still waiting for the thermal cycling result. So far I can say that it worked great on air."

So he did succeed. He then goes on to say -

"I sent a soldered 6700K to my friend Splave who is a famous overclocker in the US. Unfortunately the first LN2 test failed. At the moment I’m trying to figure out where exactly the solder preform failed – will keep you updated!"

But oddly enough there was no update. He also said, amongst that -

"During Prime95 AVX Load the CPU core temperature is at about 50 °C. The CPU was cooled by a Prolimatech Megahalems with one fan at an ambient temperature of 24 °C. Compared to stock, the temperature dropped by about 18 °C. The result is pretty similar compared to liquid metal thermal compounds. However, liquid metal thermal compounds are not working subzero that’s why I’m trying to find a way to solder Skylake."

So it definitely worked. So, as per the last debate I got involved in over this I will ask you the same questions that I asked some one else (yet they refused to answer).

1. Why can AMD solder Ryzen and why can Intel solder Xeons from the same family?
2. If Der8auer can do it somehow then don't you think Intel can too?

Edit - Rich I could not agree more, mate. I termed the phrase "Nickel and diming" which maybe was a bit off for the U.K audience but it's what the yanks refer to as basically counting beans. IE - screwing every little penny and so on wherever you can to make more yummy nummy profits.

Gawd, I bet Intel were well f****d off when AMD came along and stopped them from selling a dual core with HT with no solder on a thin substrate for £185.

 

The die size hasn't shrunk considerably since Pentium 4 days, so the small die theory doesn't hold up. Your motherboard will give out before microcracks cause a CPU to be inoperable.

Der8's method is very flawed but like you say even still 'feasible': he's not using an industrial wave solder machine that gradually elevates the temperature; it doesn't have the fine process control of specific metal depositions and LN2ing it exacerbates the metal contraction difference far beyond the normal ~30C-90C range CPUs run at.

The issues are not physical, but commercial. The indium and gold the article highlights as being rare and expensive are the issue to Intel's bottom line. Int could have R&Dd an alternative I'm sure, it's full of very clever people, but thermal paste is cheap and easy and 'good enough'. Plus if you have to de-lid that's a forced warranty void to do any real OCing it's a double-win.

65% ASP across the board is highest in the industry of chip makers I know of. Cha-ching.

 

The thing with a de-lid, for anyone knowing what they are doing ... you can get it back to 'stock' and Intel are none the wiser. However, as you've said, it basically stops unwanted OCs without voiding the warranty, and most people who do de-lid their chips will not know how to seal it up properly once the process is finished.

 

He got a chip that failed after a single thermal cycle. Production chips need to be robust enough to sell. Der8auer can cool a chip on LN2 too, so why don't Intel ship LN2? Because you can't ship chips the user needs to baby.

And Intel do NOT solder Xeons from the same family (E3 for desktop chips). I've told you this already. It's been the case since Ivy Bridge.

 

I wonder what Samsung charges and makes of Apple for use of its fabs. More than 65% profit would be a fair assumption.

65% sounds wonderful in some industry’s but it’s not actually that high in others. Samsung s8 break down even with gross over estimating had the phone around the $300 mark Samsung sells it for near £800. That’s a fun mark up over a 100% Apple is similar.

If people are daft enough to pay £185 for a duel core chip that’s there loss I certainly would not of done. It’s like paying £1000 for a phone crazy stuff.

 

It didn't fail it just wouldn't work under LN2. To which he had no answer, and he has not updated the article since. He said he was going to test the solder etc after thermal cycling but he never did.

 

I doubt they ever would but they could easily stop that by delidding the CPU themselves and seeing if it has been changed.

Yup I agree completely. I mean, they must have it bad if they make the CPU substrate thinner to save a penny here and there. It's crazy. I really couldn't believe it when Skylakes were bending under Scythe coolers and a few others. Just mental !

That said to some Intel can do no wrong. They have actually broken pretty much every law they can, used their lawyers to find loopholes in the law, paid fines for doing bad s**t because it was cheaper than paying not to do bad s**t etc.

Evil soulless money making machine.

 

Microcracks wouldn't cause a failure they'd just decrease the thermal efficiency.

Their choice to stop using solder wasn't down to a single factor, voiding and microcracks were probably the main reason but there's also things like Intel wanting to go "green" with the removal of Pb-based solders and halogen free substrate, that they had very little competition so could afford to sacrifice a few hundred Mhz, and probably least of all, because they spent a lot of money on R&D to go "green", costs.

 

Calm down dear

They sell a chip that works within the parameters they have defined, they don't need to do more, annoying as a consumer as we know it is very capable silicon.

It's business, it's not personal, companies I have worked for have spent months of man effort to save a couple of cents here or there and keep the margins above 50%, when you sell millions of units it matters, your successful cash cow keeps people in work, doing stuff you may never see, that may never generate any revenue for possible future gain that may keep you one step ahead.

Without some money behind you, there would be no progress, I can't remember exactly what the cost of my seat in the office was, it was mentioned in a prior company, but it was a few hundred thousand, due to software/hardware licensing etc, my salary was barely double digit percentage of it, multiply that by a few thousand others doing similar around the company, it all adds up, Intel is a heavy weight with its fingers in many pies, some of those pies won't be making money nor may not have for years.

 

I am perfectly calm and it was merely a statement of fact.



When you get a spare 30 mins or so sit down and watch that and make sure you understand it all. Hilariously I didn't even need to watch it, as like I mentioned elsewhere yesterday I was on the receiving end of it when the Intel rep came around and issued threats that if we dared to stock any of AMD's new Athlons they would pull our "Approved reseller" status and we would get no more Intel CPUs nor POS stuff to put in our shop windows/displays etc.