Intels Skylake-X and Coffee Lake CPUs Discussion

[Life-Binder]

http://www.gamersnexus.net/news-pc/2...-thermal-paste

[Fascinate]

SMD components sit close to the outer glue of the IHS, which means that delidding could potentially rip one of the SMDs off of the substrate

Have fun with that

Seriously, i wont buy another intel CPU unless they use solder again. 2500k had it, and it was regarded as the best CPU they ever made (as well as 2600/2700k). They charge people extra to be able to OC their CPU's, you think the least they could do is use solder on those models.

[mojo6912]

so the new intel cpus will have a chip on it to detect if it has been overclocked so they can deny warranties?

[Zenny]

so the new intel cpus will have a chip on it to detect if it has been overclocked so they can deny warranties?

I can't find any news on this? Why would they?

[mojo6912]

it says so in the article life binder posted.

"There appears to be an RFID chip in the corner of the Intel i9-7900X that we looked at, which would lead us to believe that the chip is capable of storing user information. Der8auer’s theory is that this could be used to store user overclock data, e.g. maximum stable OC. Such a chip could also be used for RMA processes, theoretically."

*puts on tinfoil hat*

why would they? to save a buttload of money of course.

[Life-Binder]

confirms 6C Coffee in August/September

[Fascinate]

Coffee lake is really what people should be excited for, x299 is just too expensive for not much gain over what AMD is offering. If coffee lake is actually 30% faster than kaby and the 6 core overclocks well...id even consider getting it.

Thing is i dont think that 30% number can be accurate being that its still on 14nm.

[Evildeffy]

Coffee lake is really what people should be excited for, x299 is just too expensive for not much gain over what AMD is offering. If coffee lake is actually 30% faster than kaby and the 6 core overclocks well...id even consider getting it.

Thing is i dont think that 30% number can be accurate being that its still on 14nm.

The advantage over Kaby Lake = Core Count.

50% extra cores, lower clocked, 30% gain.

Which means that instead of 4.5GHz turbo it'll be: probably be closer to 4.0GHz napkin math to account for both core increase and frequency decrease.
Considering it's the same architecture but optimized, you'll never in your life see a 30% uplift there.

[Fascinate]

The advantage over Kaby Lake = Core Count.

50% extra cores, lower clocked, 30% gain.

Which means that instead of 4.5GHz turbo it'll be: probably be closer to 4.0GHz napkin math to account for both core increase and frequency decrease.
Considering it's the same architecture but optimized, you'll never in your life see a 30% uplift there.

I don't think that is where intel is getting their 30% figure from, but if it is that is lol worthy. Pretty sure they are talking core per core.

[Evildeffy]

I don't think that is where intel is getting their 30% figure from, but if it is that is lol worthy. Pretty sure they are talking core per core.

Intel counts it as an entire package, they did the same with Skylake -> Kaby Lake in regards to the "just under 15% performance increase" which is of course done by a company they own which used to be an independent benchmarker.

Until at some stage, to check for fraudulent scores, the ID of the processor was changed to fool the software into thinking it was an Intel CPU instead of a VIA or AMD CPU and it magically scored considerably better.

Don't ever hold value to that stock.

Or let me phrase it this way for you:
Do you think the i7-7700K is 15% faster than the i7-6700K at the same clock speeds?

[Fascinate]

That's some pretty deceptive shit if they are comparing a 6 core coffee to a kaby quad...but you are probably right sadly.

[Life-Binder]

with 14++ nm we should probably expect at least 4.2 stock boost on 8700K, maybe it will even let them push 4.3-4.4 .. OC into 4.6-4.7 territory

yep, 7700Ks days as the reigning gaming king will soon come to an end even if there is 0% IPC increase, the cores alone will make the difference in newer games, while not losing much clocks compared to quads

for WoW a max OC 7700K will probably still remain the best



what Im wondering is if SK-X new cache system (1MB L2 per core) will be used for Coffee or not .. if it beneficial for gaming then it might

[Evildeffy]

with 14++ nm we should probably expect at least 4.2 stock boost on 8700K, maybe it will even let them push 4.3-4.4 .. OC into 4.6-4.7+ territory

yep, 7700Ks days as the reigning gaming king will soon come to an end even if there is 0% IPC increase, the cores alone will make the difference in newer games, while not losing much clocks compared to quads

What have I told you before about expectations?

In reality you're unlikely to get 4,2GHz all-core turbo, 1 or 2 cores probably yes but not all 6 as stock delivery.
It's still the same lithography and they have a power envelope to deal with.

Don't expect anything different than a Kaby Lake 7700K but with 6 cores and a lower top-out clocks in manual OCs.

New step complexity is always slower than matured previous step complexity.

what Im wondering is if SK-X new cache system (1MB L2 per core) will be used for Coffee or not .. if it beneficial for gaming then it might

It will not, again Gamers are a speck of dust in a large desert, we are simply more vocal.
It's also CONSIDERATELY more expensive to pile on the L2 cache, it would immediately increase the price by 10% if not more.

L2 cache is NOT cheap.

[Life-Binder]

will also be interesting to see OCed 8700K go up against OCed 7800X in games


8700K will most likely be clocked higher out of the box (especially since 7800X clock/boost looks to be gimped compared to 7820X), but in OC vs OC it may be equal or close

[Zenny]

The advantage over Kaby Lake = Core Count.

50% extra cores, lower clocked, 30% gain.

Which means that instead of 4.5GHz turbo it'll be: probably be closer to 4.0GHz napkin math to account for both core increase and frequency decrease.
Considering it's the same architecture but optimized, you'll never in your life see a 30% uplift there.

It's a comparison of the 15w parts from each generation, both are four core processors. It might be a mixture of clock speed increase or architectural changes, either way we are looking at a 30% performance boost at the same TDP.

[Evildeffy]

It's a comparison of the 15w parts from each generation, both are four core processors. It might be a mixture of clock speed increase or architectural changes, either way we are looking at a 30% performance boost at the same TDP.

You're looking at the Cannon Lake comparison.

They draw the same comparisons between the higher model desktop points.

If that includes TDP then it's even easier to get an item that could potentially perform far worse as it'd be a power/performance comparison only.

[pansertjald]

Damnn i found my new mobo if i go the x299 way GIGABYTE Unveils the X299 AORUS Gaming Motherboards

[Life-Binder]

https://www.pcgamesn.com/intel/intel-skylake-x-solder

Asus confirmed to us the news of X299 chips’ lack of solder is indeed true. We spoke to one of their reps during their motherboard workshop at Computex yesterday about the overclocking potential of a delidded Skylake-X CPU. They've been seeing a 30°C drop in temperatures by popping the top on the chip and using a liquid metal thermal interface.

He went on to explain Skylake-X shares the same issue as Kaby Lake - temperatures rise at lightning speed when upping the voltage going into the processor, meaning 1.3V of power is roughly the maximum amount of juice you’ll want to pump into it without serious cooling. The Asus team did manage to get 4.3GHz out of the ten-core 7900X with 1.25V of power though, so the chips will still be pretty capable even at default voltages.

so yeah good cooling required


on the other hand if even a 10c can do 4.3Ghz with just 1.25V .. and "safe" voltage with no delidding is ~1.3V

how much GHz would you extract out of a 7820X or 7800X @ 1.29-1.3V ? seems to me like 4.5+ is a safe bet, maybe even that 4.6-4.7 on the 6c


delidders with watercooling would push these chips to insane overclocks doing 1.35V+, maybe even 1.4V

- - - Updated - - -

could be golden sample though

[Gray_Matter]

There is a bit of a scathing overview here of the X-series:

http://semiaccurate.com/2017/05/30/i...thout-details/

Intel is saying in no uncertain terms that their highest end gaming part, the X which stands for Extreme series, can’t have a full bandwidth second GPU attached unless you buy a $999 part! If you think the company cares one bit about their gaming customers, this should shatter those illusions. A high-end gamer part that can only support one GPU at full bandwidth? What can I say, this is so sad it isn’t worth mocking.

[Life-Binder]

SLI/CF is dead


the fastest gaming CPU will likely end up being 8700K anyway and that supports ~16/20 lanes ? just enough for 1x Volta Titan V