Discussion Intel current and future Lakes & Rapids thread

Page 186 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
M

mikk

Diamond Member
May 15, 2012
4,133
2,136
136
RKL-S is fully 14nm so of course Gen12 is 14nm and it's most likely monolithic, Intel confirmed btw Xe is not bound to a specific process.
 
J

jpiniero

Lifer
Oct 1, 2010
14,584
5,206
136
uzzi38 said:
You know what, I'd be interested to hear what you lot have to think as to why
such a backport would have been conceived. I can only think of one reason why it would have been, and I really hope it's not the case.
Click to expand...

I'd argue that the reason Palm/Sunny Cove didn't get backported was because manufacturing convinced management that 10 nm was going to be fixable.
 
B

beginner99

Diamond Member
Jun 2, 2009
5,210
1,580
136
uzzi38 said:
You know what, I'd be interested to hear what you lot have to think as to why such a backport would have been conceived. I can only think of one reason why it would have been, and I really hope it's not the case.
Click to expand...

Why? Because 10nm is broken and staying on 14nm skylakes till 2022 or even 2023 was clearly not a sustainable route. An this still assumes 7nm doesn't fail as well. In fall 2018 intel specifically highlighted the fact that willow cove is uarch is process decoupled (in contrast to sunny cove). Only reason to mention that is because a backport was already planned.

Why are we getting comet lake on laptops? because they can't make >4 core dies on 10nm for consumer products because yields must be terrible.
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
beginner99 said:
Why? Because 10nm is broken and staying on 14nm skylakes till 2022 or even 2023 was clearly not a sustainable route. An this still assumes 7nm doesn't fail as well. In fall 2018 intel specifically highlighted the fact that willow cove is uarch is process decoupled (in contrast to sunny cove). Only reason to mention that is because a backport was already planned.

Why are we getting comet lake on laptops? because they can't make >4 core dies on 10nm for consumer products because yields must be terrible.
Click to expand...

Intel did not highlight that. They said future uArch's they'll be working on will be, they did not specify Willow Cove by name. By which point, you can be certain Intel had already begin development of Willow Cove and were probably getting ready to finalise it in under a year. The time-frames don't add up again.

And you've pretty much picked up on why I don't want Rocket Lake to be a backport. It being a backport lends itself to the idea that they think 7nm won't be in a good state at launch, and use of 14nm could go on longer than we'd all like.

That's not a situation I want Intel to be in, because it'll be a really bad place to be in in two years.
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
jpiniero said:
I'd argue that the reason Palm/Sunny Cove didn't get backported was because manufacturing convinced management that 10 nm was going to be fixable.
Click to expand...
No, it's because backports are costly as all hell and they felt no need to do so.
 
inf64

inf64

Diamond Member
Mar 11, 2011
3,697
4,015
136
Intel cannot be on 14nm Willow cove core in 1.5-2 years. That would be horrible for them and the market in general. That product would be absolutely destroyed by Zen4 on 5nm. I think they will have a 7nm Golden Cove limited launch in 2 years from now, just about enough for laptop market. I expect Willow Cove to be on 10nm but not a high core count product (max. 8 cores on mainstream and high performance ) .
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
  • Like
Reactions: lightmanek
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
jpiniero said:
They need new products to sell from now until 7 nm ships. Best case for 7 nm is mid-late 2022 at the earliest.
Click to expand...
As they've done since Sandy Bridge.
The question is, when the original plans for Rocket Lake were drafted up, where they planning on another marginal upgrade over Skylake like they've done for desktop generation after generation or did they want something substantial like Sunny Cove.

I don't see why the latter would be a priority for them.
 
Last edited:
F

FriedMoose

Member
Dec 14, 2019
48
28
51
uzzi38 said:

TGL-U. 2.3GHz base, 4.2GHz boost it seems. The score is with DDR4.

Another result that is about 10% ahead of the 1065G7 in both clock speed and single core results.

Also Gen12 graphics too:

Intel Corporation Tiger Lake Client Platform - Geekbench

Benchmark results for an Intel Corporation Tiger Lake Client Platform with an Intel 0000 processor.
browser.geekbench.com
Click to expand...
Seems like the very best 1065G7 does around 1300-1350 in GB5 windows:

1065g7 windows - Geekbench 5 CPU Search - Geekbench

browser.geekbench.com
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
FriedMoose said:
Seems like the very best 1065G7 does around 1300-1350 in GB5 windows:

1065g7 windows - Geekbench 5 CPU Search - Geekbench

browser.geekbench.com
Click to expand...
Yep. And this boosts about 300mhz higher, which is why I said approximately 10% clock bump and 10% perf bump.

Still A0 silicon, but really, by this time shouldn't we be seeing initial silicon if there's another stepping planning on being used. Especially not initial silicon that's suddenly boosting not to 3.7GHz but now 4.2GHz.

I'm starting to think TGL-U is shipping as A0.
 
T

TheGiant

Senior member
Jun 12, 2017
748
353
106
uzzi38 said:

TGL-U. 2.3GHz base, 4.2GHz boost it seems. The score is with DDR4.

Another result that is about 10% ahead of the 1065G7 in both clock speed and single core results.

Also Gen12 graphics too:

Intel Corporation Tiger Lake Client Platform - Geekbench

Benchmark results for an Intel Corporation Tiger Lake Client Platform with an Intel 0000 processor.
browser.geekbench.com
Click to expand...
looks to me like TGL has low IPC increase

but the clock is nice for U version- macbook air with this and the new gen12 gpu look nice
 
DrMrLordX

DrMrLordX

Lifer
Apr 27, 2000
21,617
10,826
136
beginner99 said:
That can only work if has pcie4 and is willow-cove based.
Click to expand...

Rocket Lake-S will likely enjoy PCIe4 support, as will Alder Lake-S. I think it was agreed that Rocket Lake-S would stay on LGA1200 (with a new chipset) while Alder Lake-S will be on LGA1700. Also, it looks like Rocket Lake will be based on Willow Cove.

inf64 said:
I think they will have a 7nm Golden Cove limited launch in 2 years from now, just about enough for laptop market.
Click to expand...

Allegedly, Alder Lake-P and Alder Lake-S will be 10nm Golden Cove, released maybe as soon as one quarter after Rocket Lake-S hits the market. How they'll coexist isn't clear, though availability of these parts may be very low. All this happens in 2021. What isn't clear is when exactly Rocket Lake-S shows up: is it Q1 or Q2 2021? Alder Lake-S looks like Q3 2021, if it ever appears.
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
DrMrLordX said:
What isn't clear is when exactly Rocket Lake-S shows up: is it Q1 or Q2 2021? Alder Lake-S looks like Q3 2021, if it ever appears.
Click to expand...
I'd say about then for RKL-S. Either Q1 or Q2.

I don't see ADL-S landing till 2022. Another launch in 2021 would leave nothing for 2022, as I doubt we'll see 7nm desktop until late 2023
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
FriedMoose said:
Now that I think about it, it's probably better to compare vs another ES. If you look at GB scores, ES are almost never even close to the best even if they have retail clocks.
Click to expand...
That's usually due to poor boost clocks, or god awful memory configurations. That latter usually happens when the CPU ends up in the hands of OEMs.

This score is already significantly above ICL almost across the board. Compare the Integer and Floating point scores - it's quite a decent uplift. But you shouldn't start expecting huge uplifts from here - we're 6 months away from a launch, at most we'll probably see clocks improved slightly.
 
uzzi38

uzzi38

Platinum Member
Oct 16, 2019
2,622
5,880
146
DrMrLordX said:
Intel better hope that isn't true. Aurora is supposed to be ready in Q4 2021 with other 7nm coming in 2022 . . . that's their lifeline.
Click to expand...
...I'm sorry, I'm an idiot. I've mixed up my years, you're correct.

In which case, I'll update to Q1 2022 for ADL-S and Q4 MTL-S I guess?
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom