Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads

[Tigerick]

Wildcat Lake (WCL) Preliminary Specs

Intel Wildcat Lake (WCL) is upcoming mobile SoC replacing ADL-N. WCL consists of 2 tiles: compute tile and PCD tile. It is true single die consists of CPU, GPU and NPU that is fabbed by 18-A process. Last time I checked, PCD tile is fabbed by TSMC N6 process. They are connected through UCIe, not D2D; a first from Intel. Expecting launching in Q2/Computex 2026. In case people don't remember AlderLake-N, I have created a table below to compare the detail specs of ADL-N and WCL. Just for fun, I am throwing LNL and upcoming Mediatek D9500 SoC.

		Intel Alder Lake - N	Intel Wildcat Lake	Intel Lunar Lake		Mediatek D9500
Launch Date		Q1-2023	Q2-2026 ?	Q3-2024		Q3-2025
Model		Intel N300	?	Core Ultra 7 268V		Dimensity 9500 5G
Dies		2	2	2		1
Node		Intel 7 + ?	Intel 18-A + TSMC N6	TSMC N3B + N6		TSMC N3P
CPU		8 E-cores	2 P-core + 4 LP E-cores	4 P-core + 4 LP E-cores		C1 1+3+4
Threads		8	6	8		8
Max Clock		3.8 GHz	?	5 GHz
L3 Cache		6 MB	?	12 MB
TDP		7 W	Fanless ?	17 W		Fanless
Memory		64-bit LPDDR5-4800	64-bit LPDDR5-6800 ?	128-bit LPDDR5X-8533		64-bit LPDDR5X-10667
Size		16 GB	?	32 GB		24 GB ?
Bandwidth			~ 55 GB/s	136 GB/s		85.6 GB/s
GPU		UHD Graphics		Arc 140V		G1 Ultra
EU / Xe		32 EU	2 Xe	8 Xe		12
Max Clock		1.25 GHz		2 GHz
NPU		NA	18 TOPS	48 TOPS		100 TOPS ?

As Hot Chips 34 starting this week, Intel will unveil technical information of upcoming Meteor Lake (MTL) and Arrow Lake (ARL), new generation platform after Raptor Lake. Both MTL and ARL represent new direction which Intel will move to multiple chiplets and combine as one SoC platform.

MTL also represents new compute tile that based on Intel 4 process which is based on EUV lithography, a first from Intel. Intel expects to ship MTL mobile SoC in 2023.

ARL will come after MTL so Intel should be shipping it in 2024, that is what Intel roadmap is telling us. ARL compute tile will be manufactured by Intel 20A process, a first from Intel to use GAA transistors called RibbonFET.

 

[H433x0n]

You know David Huang says it cut uOP cache sizes to Zen 3 levels and even PRF sizes right? Hence the 16% gains despite big changes elsewhere.

The core is ~35% bigger (with same amount of L2$) while using a node that's a 0.94 shrink. It’s not as if it’s some small changes and a minor increase in resources.

 

[SiliconFly]

-

 

[yuri69]

The core is ~35% bigger (with same amount of L2$) while using a node that's a 0.94 shrink. It’s not as if it’s some small changes and a minor increase in resources.

The core got bloated by that very generous AVX512 implementation, didn't it?

 

[DrMrLordX]

You were the one complaining that the leaked slide last year was optimistic

That wasn't exactly a complaint, it was more a lament. Though 3%/15% is still pretty close to the slide. The point is: the performance increases shown by the old slide were kinda crap, so an ES2 (which we now know to have been mostly unlocked) reaching even lower performance improvements is not a good thing.

I'm still wondering why you even bothered with this line of conversation at all.

 

[coercitiv]

(Late) October is back.

 

[cannedlake240]

They will be replaced.

Apparently not. If RYC cancellation rumors are of any substance, Core is their main architecture for the foreseeable future, unless atom transforms into a P core

 

[majord]

No way man. Desktop on 3nm TSM should clock every bit as good as 5/4nm TSM. Surely its not hitting 6.2, but I'd put big money that its likely to clock right up there with Zen 4 and 5. 5.6-5.7 for sure, and maybe more.

it's not all about the process' raw capabilities

This is Intel's first performance core on TSMC IIRC , first P core ported to any other process but their own, it remains to be seen if they can pull decent peak clocks , without the sort of "legacy" in house process tuning they relied on for yrs.

I'm quite surprised Lion cove isn't shooting for lower clock higher IPC than it is .. Surprised even AMD are still shooting for relatively high clock speeds , but they've been at this for a while (since KAveri/Excavator basically)

 

[deasd]

Another LNL GB result

https://twitter.com/x/status/1811661027064287393

(Late) October is back.

View attachment 102864

Benchlife has reiterated before that ARL-S in October last month. But from Jaykihn we know the QS is also in October, which raise doubt. Jaykihn also said this time is different, very short gap between QS and hit shelves.

MLID has old video stated QS in October and he was not confident that it could be launch before the end of this year, that video is 3-4 months old but I cannot find the video back.

 

[whoshere]

Newest Intel(R) Core(TM) Ultra 7 256V GB6 results:



Frequencies:

0 4763
1 4615
2 4731
3 4776
4 4777
5 4773
6 4748
7 4773
8 4782
9 4778
10 4764
11 4759
12 4774
13 4579
14 4754
15 4767
16 4658
17 4694
18 4776
19 4755
20 4719
21 4783
22 4738
23 4775
24 4779
25 4765
26 4783
27 4778
28 4760
29 4767
30 4774
31 4775
32 4773
33 4758
34 4697
35 4768
36 4557
37 4423
38 4745
39 4736
40 4754
41 4724
42 4715
43 4582
44 4721
45 4654
46 4698
47 4739

Matches 1.5 year old Ryzen 7 7840HS in ST mode.

 

[Det0x]

Newest Intel(R) Core(TM) Ultra 7 256V GB6 results:

View attachment 102873

Frequencies:

0 4763
1 4615
2 4731
3 4776
4 4777
5 4773
6 4748
7 4773
8 4782
9 4778
10 4764
11 4759
12 4774
13 4579
14 4754
15 4767
16 4658
17 4694
18 4776
19 4755
20 4719
21 4783
22 4738
23 4775
24 4779
25 4765
26 4783
27 4778
28 4760
29 4767
30 4774
31 4775
32 4773
33 4758
34 4697
35 4768
36 4557
37 4423
38 4745
39 4736
40 4754
41 4724
42 4715
43 4582
44 4721
45 4654
46 4698
47 4739

Matches 1.5 year old Ryzen 7 7840HS in ST mode.

GB 6.0 scores ~10%lower MT than 6.1/6.2 and 6.3

 

[cebri1]

Newest Intel(R) Core(TM) Ultra 7 256V GB6 results:

Matches 1.5 year old Ryzen 7 7840HS in ST mode.

30W vs 54W

 

[FlameTail]

Newest Intel(R) Core(TM) Ultra 7 256V GB6 results:

View attachment 102873

Frequencies:

ST and MT scores are essentially identical to Apple M2

 

[poke01]

the ultra 9 lunar will match M3?

 

[SiliconFly]

-

 

[poke01]

Doubtful

It might cause it has PL2 of 30watts

 

[FlameTail]

They are not going to be feeding 30W into a single core...

 

[inf64]

2600 pts for ST at 4.7Ghz boost. Compare to Z5 @ 5.7Ghz and 3400pts. Z5 around 8-9% faster per clock than Lion Cove version in LNL in this version of GB

 

[SiliconFly]

-

 

[Tup3x]

2600 pts for ST at 4.7Ghz boost. Compare to Z5 @ 5.7Ghz and 3400pts. Z5 around 8-9% faster per clock than Lion Cove version in LNL in this version of GB

Zen5 scores aren't from GB 6.0 though? Not comparable.

 

[DavidC1]

The core is ~35% bigger (with same amount of L2$) while using a node that's a 0.94 shrink. It’s not as if it’s some small changes and a minor increase in resources.

You have the overkill AVX512 implementation too.

And the biggest evidence is that it's only 16% gain. So it suggests it's not a big of a gain as core size increase, unless they did something else like focus on more clocks or something.

 

[inf64]

Zen5 scores aren't from GB 6.0 though? Not comparable.

Yes they are:

AMD Ryzen 9 9950X flagship Zen5 CPU spotted on Geekbench, new 9900X/9700X/9600X results emerge - VideoCardz.com

AMD Ryzen 9000 in Geekbench New benches were made public, featuring the very first score of 16-core variant. It looks like this might be the final update on Geekbench scores as we now have a complete lineup in one chart, featuring the first results for the Ryzen 9 9950X and updated scores for...

videocardz.com

Edit: Oh you meant the subversion is different? That is not a big deal I think.

 

[Hitman928]

Yes they are:

AMD Ryzen 9 9950X flagship Zen5 CPU spotted on Geekbench, new 9900X/9700X/9600X results emerge - VideoCardz.com

AMD Ryzen 9000 in Geekbench New benches were made public, featuring the very first score of 16-core variant. It looks like this might be the final update on Geekbench scores as we now have a complete lineup in one chart, featuring the first results for the Ryzen 9 9950X and updated scores for...

videocardz.com

Edit: Oh you meant the subversion is different? That is not a big deal I think.

See here:

Geekbench 6.1 - Geekbench Blog

www.geekbench.com

How much a difference it makes for LNL is unknown.

 

[inf64]

See here:

Geekbench 6.1 - Geekbench Blog

www.geekbench.com

How much a difference it makes for LNL is unknown.

Thank you very much, I was not aware of this change.

 

[H433x0n]

2600 pts for ST at 4.7Ghz boost. Compare to Z5 @ 5.7Ghz and 3400pts. Z5 around 8-9% faster per clock than Lion Cove version in LNL in this version of GB

You’re comparing a mobile core in a low end mobile SKU against the best score ever recorded for a Zen 5 desktop CPU. Most of the Zen 5 GB6 results are in the 3250-3350 range.

I could cherry pick a low scoring Strix sample and make the opposite argument. We should probably wait for data on Lion Cove on desktop before making determinations.

 

[DavidC1]

2600 pts for ST at 4.7Ghz boost. Compare to Z5 @ 5.7Ghz and 3400pts. Z5 around 8-9% faster per clock than Lion Cove version in LNL in this version of GB

Raptorlake desktop is 7% faster than Raptorlake mobile, which will make up for the difference.

Consider that it'll run laps around Zen 5 mobile for battery life, so even if it had 7% deficit, no one buying it will care.