Coffeelake thread, benchmarks, reviews, input, everything.

[JoeRambo]

This is getting beyond hilarious, Intel seems to be

By that first pic though, I could conclude that the 270 board only supports 3 cores.

Or "4" are implied by "ALL setting, one that does not limit anything. And they chose to spell out 8 for new gen chipset.

 

[LTC8K6]

Seems odd that 370 and 310 have both ALL and 8 as choices, whereas 270 seems to properly have ALL and 3, ALL meaning 4.

7 should be the last choice for 370/310, as all means 8.

 

[Dayman1225]

They appear to be real thanks to a reddit user by the name of eric98k providing the tools to do so

Here are my results in screenshots:

Nice find, @dooon

edit:

eric98k also tested the Z370 Asrock Pro 4 and found similar results

He also mentions in his reddit comment that these options existed from the first Z370 bios'

 

[mikk]

Obviously the power delivery change from last year was made with 8 cores in mind.

 

[utahraptor]

According to benchlife the original Z390 chipset (14nm with native USB 3.1 Gen2) might be canceled, instead they will rename Z370 to Z390. Main reason seems to be the 14nm capacity, because of the 10nm delays Intels 14nm demand is too high at this point. So the only difference might be the power delivery from the socket, it's unclear if they have to change it once again to support the 8 Core version.

https://benchlife.info/intel-might-rename-22nm-z370-to-replace-14nm-z390-chipset-06222018/

Here is more support to Z390 being a re-badge: Z390 chipset now rumored to get rebranded from Z370

 

[jpiniero]

https://videocardz.com/76709/intel-officially-confirms-core-9000-series-based-on-coffee-lake-s

In the Spectre/Meltdown microcode document, Intel mentions 9th gen Coffee Lake. Not exactly new news but basically a confirmation. No mention of any i7 models, just the i5 and i3.

 

[LTC8K6]

So it looks like Wikichip had the lineup correct all along, and the 8 core will likely be called 9700K.

https://en.wikichip.org/wiki/intel/cores/coffee_lake_r#Coffee_Lake_R_Processors

 

[mikk]

The number of threads might be incorrect.

https://twitter.com/momomo_us/status/1014149205957304320

i5-9600K with 9BM L3, I would expect 12 MB if it would come with 12 Threads. And there is a Core i9 with 8C tbd, so possibly it will be named i9-9700K.

 

[LTC8K6]

The number of threads might be incorrect.

https://twitter.com/momomo_us/status/1014149205957304320

i5-9600K with 9BM L3, I would expect 12 MB if it would come with 12 Threads. And there is a Core i9 with 8C tbd, so possibly it will be named i9-9700K.

9700K 6/12
9900k 8/16

?

 

[dooon]

yummy.

8th Generation Intel Core Processor Spec Update

 

[LTC8K6]

8650K and 9600K appear to be identical.

I see a few pairs like that.

 

[PeterScott]

8650K and 9600K appear to be identical.

I see a few pairs like that.

Is there any way to tell if the new 9xxx chip designations have HT enabled from available info.

 

[LTC8K6]

I think it's the same deal.

i3 no 4/4
i5 no 6/6
i7/i9 yes 6/12 8/16

But it seems like there may be some 4/8 chips in there.

 

[EXCellR8]

The full 9000 series lineup is likely to be officially introduced with 8-core SKUs, which are not present in the same documents yet.

Interestingly, the 9th Gen Core series are listed as 8th Gen Core parts, which might be found confusing.

Source

 

[LTC8K6]

They are Coffee Lake, so 8th gen.

 

[PeterScott]

So nothing but re-brands so far (100 MHz clock speed bump is insignificant). For 9xxx, it is likely the only new die will be 8 core version.

The stagnation at Intel is just baffling. Year after year of *Lake cores that have no significant IPC improvements along with year after year of 14nm process repeats.

This is almost what you would expect if Intel deprecated x86 CPUs, to focus on something bigger, except x86 is by far the biggest, and most important product for Intel.

IMO BK should have been openly fired for gross incompetence, for allowing this to happen. Intel has the resources to have better contingency plans than essentially nothing. While more cores are nice, they don't hide the fact that IPC and Process have both been stagnant for years.

 

[jpiniero]

I was thinking perhaps there will be 35/65 W 8C locked i7's with the cache cut to 12 MB to save power... and the K model is branded i9 because they are going to charge more (eg: $425)

 

[jpiniero]

Whiskey Lake spotted, i5 8265U (1.6/3.9) and i7 8565U (1.8/4.5).

https://imgur.com/oikI0Di
https://translate.google.com/transl...hangyongbijiben/201806/t20180621_407541.shtml

 

[Jan Olšan]

8650K and 9600K appear to be identical.

I see a few pairs like that.

What seems to have happened: those matching CPUs (8650, 8120, 8420 etc) were how Intel originally wanted to introduce this Coffee Lake refresh, just as a simple mid-cycle speedbump of 8000 gen. They changed mind, cancelled those and made it generation 9000. That's why the whole document is called spec update. These Intel docs announce changes done to specific SKUs, which means that for example i5-9600K doesn't actually have the same parameters as i5-8650K, IT IS THE ACTUAL SAME SKU.

The important takeaway from that is that most likely none of these chips will have HT enabled. That would have been too big a change for a speedbump refresh inside 8000 line. And the 9000 chips will have exactly the same parameters, it follows from the "spec update" angle.

 

[LTC8K6]

I was thinking 8650K might work in older boards, 9600K in newer?

 

[epsilon84]

So nothing but re-brands so far (100 MHz clock speed bump is insignificant). For 9xxx, it is likely the only new die will be 8 core version.

The stagnation at Intel is just baffling. Year after year of *Lake cores that have no significant IPC improvements along with year after year of 14nm process repeats.

This is almost what you would expect if Intel deprecated x86 CPUs, to focus on something bigger, except x86 is by far the biggest, and most important product for Intel.

IMO BK should have been openly fired for gross incompetence, for allowing this to happen. Intel has the resources to have better contingency plans than essentially nothing. While more cores are nice, they don't hide the fact that IPC and Process have both been stagnant for years.

Well when the 10nm process is broken and massively delayed... I think a doubling of core count within a 24 month period is a reasonable compromise. Despite being stuck at 14nm for the past 3 years with no IPC gains as you said, we are actually getting much bigger gains in MT performance compared to the past, thanks in no small part to AMD releasing Ryzen of course.

 

[PeterScott]

Well when the 10nm process is broken and massively delayed... I think a doubling of core count within a 24 month period is a reasonable compromise. Despite being stuck at 14nm for the past 3 years with no IPC gains as you said, we are actually getting much bigger gains in MT performance compared to the past, thanks in no small part to AMD releasing Ryzen of course.

Which I noted. It's nice to get more cores, but it doesn't hide how stagnant intel has been on IPC/Process.

 

[frozentundra123456]

So nothing but re-brands so far (100 MHz clock speed bump is insignificant). For 9xxx, it is likely the only new die will be 8 core version.

The stagnation at Intel is just baffling. Year after year of *Lake cores that have no significant IPC improvements along with year after year of 14nm process repeats.

This is almost what you would expect if Intel deprecated x86 CPUs, to focus on something bigger, except x86 is by far the biggest, and most important product for Intel.

IMO BK should have been openly fired for gross incompetence, for allowing this to happen. Intel has the resources to have better contingency plans than essentially nothing. While more cores are nice, they don't hide the fact that IPC and Process have both been stagnant for years.

Any yet, with a brand new design and an independent foundry, AMD is slightly behind in IPC for most applications, and clearly behind in clockspeed. Maybe improvements are just not that easy anymore.

 

[PeterScott]

Any yet, with a brand new design and an independent foundry, AMD is slightly behind in IPC for most applications, and clearly behind in clockspeed. Maybe improvements are just not that easy anymore.

I am not downplaying Intels huge initial lead, and no doubt we are in an era of diminishing returns.

But Intel has pretty much had no recent improvements in architecture. Has the architecture team been doing nothing in the *Lake era?

 

[maddie]

I am not downplaying Intels huge initial lead, and no doubt we are in an era of diminishing returns.

But Intel has pretty much had no recent improvements in architecture. Has the architecture team been doing nothing in the *Lake era?

Probably any new methods leading to reasonable IPC increases needed a large transistor budget. I can also see them continually expecting that 10nm would be ready soon. If this is true, then internally Intel must be fractured.