Intel to launch 6 core Coffee Lake-S CPUs & Z370 chipset 5 October 2017

[~>Dg<~]

the motherboard thing for a start.

who realistically will be looking at the cpu we mentioning ? 95 percent wont be on z170 z270 boards.they will be on older gear.so it doesnt matter ! for 5 percent it maybe but on the whole it wont.

got to be delidded. for 95 percent it wont matter.it will clock higher than ryzen chips as it is and be fine.maybe for absolute clocks for those very small percentage its true but for the masses it will be fine.


so you have a quicker cpu for gaming and multithreaded stuff.for the masses.true it wont tick everyones box but nothing does.

 

[DragonQ]

the motherboard thing for a start.

who realistically will be looking at the cpu we mentioning ? 95 percent wont be on z170 z270 boards.they will be on older gear.so it doesnt matter ! for 5 percent it maybe but on the whole it wont.

Yes you're probably right but how long have people have been waiting for Intel to up the core count on their mainstream CPUs? I bet a lot of Skylake/Kaby Lake owners would jump at a drop-in 6c/6t or 6c/12t replacement, or would have bought them two years ago if they'd been available. Now they may not due to the extra overhead.

got to be delidded. for 95 percent it wont matter.it will clock higher than ryzen chips as it is and be fine.maybe for absolute clocks for those very small percentage its true but for the masses it will be fine.

Depends how high you want to clock it. Obviously anyone serious about clocking them (e.g. us forum users) will need either a delid or expensive cooling, or both. Most users don't overclock anyway so it's a moot point for them, but since most comparisons between Intel and AMD CPUs tend to involve the argument that Intel's chips clock to 4.9+ GHz, this has to be taken into account.

so you have a quicker cpu for gaming and multithreaded stuff.for the masses.true it wont tick everyones box but nothing does.

Yes, the i7-8700K will be fast. Will it be fast enough to justify the (likely) double cost compared to the R5 1600? Not for me and I suspect not for most people. You'd need a serious budget to even think about buying one for a gaming machine; it actually looks more reasonable for a mixed-workload machine actually. I bet the i5-8600K sells better though.

 

[~>Dg<~]

funny thing is you could have had the value 5820k from 3 years ago fast as the ryzen out now for the same price as today.so the chips have been there.

 

[SiDeards73]

Those supposed Cinebench scores for the 8700k dont exactly set the world on fire though, but pretty much sum up what i thought, its slightly slower than a 7700k in single threaded stuff and slightly slower than the 8 Core Ryzens in Multithreaded. Not a bad chip, but it doesnt make much sense to anyone who only games and is already on a 7700k tbh.

And the points about having to De-lid is kinda on point in some regards, most people buying the 8700k and even the 7700k for gaming are enthuisiasts who will likely de-lid them as they will want the best performance and cooling from them. I still find it shoddy that Intel cant actually put decent TIM in place for these types of chips

This will be my first foray into delidding, im not looking forward to it in all honesty lol

 

[Journey]

This will be my first foray into delidding, im not looking forward to it in all honesty lol

It's easy, I did a load of Opteron 146's etc. back in the day with a stanley blade, by hand. Having to invalidate your warranty, on a brand new expensive CPU, that's something you just have to live with. Make sure you test it thoroughly first!

 

[DragonQ]

funny thing is you could have had the value 5820k from 3 years ago fast as the ryzen out now for the same price as today.so the chips have been there.

Yes that chip was fantastic before it went up in price. Some people were still recommending the i7-6700K though for the same price, just for the 3-5% IPC boost! The motherboards were more expensive than AM4 ones though.

 

[~>Dg<~]

yes but factor in you could have have todays speed for 3 years plus for £100 more mobo.i know which is the better value.3 years worth of use and still going for that 100 quid.

 

[gavinh87]

I agree with the delidding comments. Not something I'd want to do, invalidating warranty for the sake of 100-200mhz, and looking at benchmarks a stock 7700k isn't exactly slow compared to a 5ghz 7700k.

 

[DragonQ]

yes but factor in you could have have todays speed for 3 years plus for £100 more mobo.i know which is the better value.3 years worth of use and still going for that 100 quid.

Hindsight is a wonderful thing. I wonder how many predicted Intel's drip-feed IPC improvements would get even slower?

 

[Hagar]

I agree with the delidding comments. Not something I'd want to do, invalidating warranty for the sake of 100-200mhz, and looking at benchmarks a stock 7700k isn't exactly slow compared to a 5ghz 7700k.

I have never fully understood the rationale behind the heat spreader anyway. I had no problems with the bare CPU of the past and adding another layer of TIM and aluminium appears to just add another issue.

 

[Journey]

I have never fully understood the rationale behind the heat spreader anyway. I had no problems with the bare CPU of the past and adding another layer of TIM and aluminium appears to just add another issue.

A lot of it was down to the damage that could be easily done to the bare die package, and with die's getting bigger, and bigger it was getting to easy to crush the edges. I think heat spreaders are sensible, as long as they are mounted to the die package with high quality materials, and not huge gaps due to the sealant being used.

 

[Hagar]

A lot of it was down to the damage that could be easily done to the bare die package, and with die's getting bigger, and bigger it was getting to easy to crush the edges. I think heat spreaders are sensible, as long as they are mounted to the die package with high quality materials, and not huge gaps due to the sealant being used.

In which case a better engineered solution would be a shim, directly exposing the die surface to the heatsink. But off topic and one for another discussion. I agree with you on material selection.

 

[SiDeards73]

My issue is with the rubbish TIM they use, if your running at stock then its probably a non issue, but if you want to squeeze the extra performance out of the chips you basically have to delid them now to get the thermals under control, i can imagine the 8700K being worse to cool than the 7700k, and plenty of videos have shown how bad their TIM is, changing it normally knocks between 10-20c off the temps, thats a lot of thermal reduction.

 

[Perfect_Chaos]

Hindsight is a wonderful thing. I wonder how many predicted Intel's drip-feed IPC improvements would get even slower?

Intel said themselves that IPC would likely regress a bit in future, i remember at least 1 article on that. Though they weren't very clear on how long that would last, surely it won't be that way with Icelake.

 

[Jono8]

those latest cinebench scores are very disappointing.

 

[Journey]

My issue is with the rubbish TIM they use, if your running at stock then its probably a non issue, but if you want to squeeze the extra performance out of the chips you basically have to delid them now to get the thermals under control, i can imagine the 8700K being worse to cool than the 7700k, and plenty of videos have shown how bad their TIM is, changing it normally knocks between 10-20c off the temps, thats a lot of thermal reduction.

It's not the TIM, it's the large gap between the die package and the heat spreader, meaning the CPU is not being cooled efficiently. This is due to the mounting of the heat spreader, and the sealant used to secure it. Removing the sealant, and putting the heat spreader back on, with the same TIM would cause similar temp drops. It's a bit like having an ill fitting HSF, where the contact is not quite good enough, it still works but it gets hotter than needs be.

 

[SiDeards73]

It's not the TIM, it's the large gap between the die package and the heat spreader, meaning the CPU is not being cooled efficiently. This is due to the mounting of the heat spreader, and the sealant used to secure it. Removing the sealant, and putting the heat spreader back on, with the same TIM would cause similar temp drops. It's a bit like having an ill fitting HSF, where the contact is not quite good enough, it still works but it gets hotter than needs be.

Interesting, so what would you propose to stick the IHS back on with if your taking all the sealant off? I am edging more towards just paying for a pre-binned CPU in all honesty, although i hate to do that, but all the hassle of speed binning and delidding etc is all done for you so your paying for that

Might just end up getting an 8Pack pre OC'd bundle tbh.

 

[Journey]

Interesting, so what would you propose to stick the IHS back on with if your taking all the sealant off?

No need to stick it on, the socket will happily hold it in place once closed. Making sure the gap is as small as possible, between the die and the spreader and therefore better heat transfer.

 

[Drollic]

Performance will be excellent, I'm sure, but it seems to me that the concerns raised in this thread are legitimate. It's more expensive than the i7-7700K, it probably won't clock as high, it still requires delidding and/or expensive cooling to get a decent overclock, and it requires yet another new motherboard. I'm sure you'll say you don't care and that but others do. Even if you think it's the best thing out there, that doesn't mean there can't be negatives about it.

Except most aren't legitimate. It's more expensive - it has an extra two cores, it won't clock as high - it has an extra two cores, it still requires decent cooling - no change here, two extra cores. Yet another new motherboard - why wouldn't Intel when the alternative isn't great unless you multitask, are budget constrained or don't game much.

 

[DragonQ]

Except most aren't legitimate. It's more expensive - it has an extra two cores

And yet AMD's chips have more cores and aren't more expensive. Intel doesn't live in a vacuum (any more).

it won't clock as high - it has an extra two cores

Sure, no-one would expect it to clock as high. Clocks are kinda important when IPC has dried up so much though and lower clocks means it'll be slightly worse than the i7-7700K in certain workloads. AMD's chips are obviously even slower in single-core performance but that doesn't mean this isn't a legitimate gripe.

it still requires decent cooling - no change here, two extra cores.

And yet AMD's chips have more cores and better temperatures. Intel doesn't live in a vacuum (any more). The combination of Intel chasing GHz and not soldering their IHSs is the problem here, the extra two cores isn't really the cause of the problem.

Yet another new motherboard - why wouldn't Intel

Huh? Saying "well obviously Intel did this because they can get away with it" doesn't make the grievance illegitimate.

the alternative isn't great unless you multitask

So pretty much everyone.

are budget constrained

So pretty much everyone.

or don't game much.

So pretty much everyone.