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.

 

[MS_AT]

Yeah, I know. But I figured it should give at least some indication of how well the CPU performs with video transcoding in general. If it performs well with video transcoding at peak performance / max TDP (as used in that test), and it has a sweet spot at around 100 W (after which more watts does not give much more perf), then I figure it's likely to have good video transcoding perf and perf/watt at 100W also.

I've not been able to find any comparison of video transcoding performance at ~100 W between 285K / 14900K / 9950X(3D). But if someone has found any such comparison, please share. Ideally for x265 video transcoding, since that's probably what I'll do most of.

Well, then outlets I was able to find quickly, if tested at normal TDPs put Ryzen ahead https://www.phoronix.com/review/amd-ryzen-9-9950x3d-linux/7 https://www.techpowerup.com/review/amd-ryzen-9-9950x3d/11.html https://www.tomshardware.com/pc-components/cpus/amd-ryzen-9-9950x3d-review/3 but it's hard for me to comment how sensible these benchmarks are, still if you were wondering about an indication. Oh and here notebookcheck has power scaling tests in Cinebench 23/24 between 9955hx and 275hx at the bottom of the page, covering the points that might interest you if only to give some +/- estimates (still as I said, they are rather weak proxies) https://www.notebookcheck.net/AMD-R...ttle-for-the-fastest-gaming-CPU.989996.0.html

For more detailed x265 benchmarks you could try to look for some dedicated x265 / encoding message boards.

 

[csbin]

NVL-S
LGA 1954

https://twitter.com/x/status/1913414611887669723

 

[alcoholbob]

Gonna be the new Cinememe champion with 52 cores.

 

[poke01]

NVL-S
LGA 1954

https://twitter.com/x/status/1913414611887669723

View attachment 122311

meh, new socket. Why?

Meanwhile, AMD will release Zen6 on AM5.

 

[alcoholbob]

Rumors are no stacked cache until the successor to Nova Lake too. Which means Nova Lake is probably gonna be Arrow Lake 2.0. Gonna get rolled in gaming and most practical workloads and will exist to win all-core benchmarks like Cinebench, and have horrible power efficiency again even with node advantage. There's probably a reason why they are going all out in core count (16P + 32E). Likely the cores themselves will be uncompetitive in IPC.

 

[511]

Rumors are no stacked cache until the successor to Nova Lake too. Which means Nova Lake is probably gonna be Arrow Lake 2.0. Gonna get rolled in gaming and most practical workloads and will exist to win all-core benchmarks like Cinebench, and have horrible power efficiency again even with node advantage. There's probably a reason why they are going all out in core count (16P + 32E). Likely the cores themselves will be uncompetitive in IPC.

False NVL has a SKU Planned with 144 MB L3 as for IPC Panther Cove is not a small upgrade and neither is Arctic Wolf I fully expect both to be 20%+ over LNC/Skymont just cause they are increasing core count doesn't mean it is weak was Alder Lake weak?

 

[LightningZ71]

It would be unlikely for them to not make any effort to address the uncore/fabric issues that Arrow Lake has as well, even if it's largely just increasing fabric clocks everywhere.

 

[jpiniero]

meh, new socket. Why?

Intel's been doing 2 gens per socket for so long. OEMs don't seem to mind.

 

[511]

meh, new socket. Why?

Meanwhile, AMD will release Zen6 on AM5.

maybe different power Delivery mechanism and electrical specs to allow for 52 Core Nova Lake

 

[fastandfurious6]

False NVL has a SKU Planned with 144 MB L3

intel 3d cache confirmed?

 

[511]

intel 3d cache confirmed?

i don't know exactly what's Intel is using for it but it is a 144MB L3 cache SKU for sure in NVL linup btw i have found the reason for Lion Cove's bad PPA it's too funny when i heard about it

 

[MS_AT]

Lion Cove's bad PPA it's too funny when i heard about it

Care to share?

but it is a 144MB L3 cache SKU

Wasn't this SKU cancelled?

 

[AMDK11]

i don't know exactly what's Intel is using for it but it is a 144MB L3 cache SKU for sure in NVL linup btw i have found the reason for Lion Cove's bad PPA it's too funny when i heard about it

Could you please explain to us the reason for the bad PPA of the LionCove core? Is there more to what we already know?

ArrowLake L3 36MB, 84 cycles(14.83 ns), 1.02 TB/s
Zen5 L3 32MB, 48 cycles(8.44 ns), 1.6 TB/s

Zen5 L3
48 cykli (8.44 ns)
1T: 173.6 GB/s
nT: 1.6 TB/s

LionCove L3
84 cykle (14.83 na)
1T: 57 GB/s
nT: 1.02 TB/s

 

[511]

It was the move to Industry standard tooling so they they messed up the physical design part as the P core team is using Industry standard tooling fir the first time.

 

[AMDK11]

Yes, I knew about the new design tools that Intel had switched to and that were used in the LionCove project, among others. But is that the main reason?

 

[511]

Yes, I knew about the new design tools that Intel had switched to and that were used in the LionCove project, among others. But is that the main reason?

Yes one of the biggest reason regarding it having horrendous PPA the L3/NoC is done by another team.

 

[AMDK11]

Yes one of the biggest reason regarding it having horrendous PPA the L3/NoC is done by another team.

So LionCove+L2 was designed by one team and L3 by another?

 

[511]

So LionCove+L2 was designed by one team and L3 by another?

Yes Lion Cove is designed by team in Haifa Skymont by Austin team and L3 by another team and NOC by another

 

[AMDK11]

LionCove has a very good L2 and a pretty good L1-1.5-D subsystem. Maybe L1.5-D could be a bit better but it's not bad. L3 is a nightmare. If the same team that did the LionCove core took on L3 it would be a lot better.

Are there any sources stating which parts of ArrowLake were designed by which team?

 

[511]

LionCove has a very good L2 and a pretty good L1-1.5-D subsystem. Maybe L1.5-D could be a bit better but it's not bad. L3 is a nightmare. If the same team that did the LionCove core took on L3 it would be a lot better.

Are there any sources stating which parts of ArrowLake were designed by which team?

The interview with the Core team by kitguru they mentioned that they only worked on the core part of the core as for the L3 I think it will be fixed by PTL and by NVL they would improve upon it.

 

[AMDK11]

The interview with the Core team by kitguru they mentioned that they only worked on the core part of the core as for the L3 I think it will be fixed by PTL and by NVL they would improve upon it.

Thanks for the reply. I'll look into that interview for Kitguru.

 

[OneEng2]

meh, new socket. Why?

Meanwhile, AMD will release Zen6 on AM5.

I wondered the same thing!

Rumors are no stacked cache until the successor to Nova Lake too. Which means Nova Lake is probably gonna be Arrow Lake 2.0. Gonna get rolled in gaming and most practical workloads and will exist to win all-core benchmarks like Cinebench, and have horrible power efficiency again even with node advantage. There's probably a reason why they are going all out in core count (16P + 32E). Likely the cores themselves will be uncompetitive in IPC.

It's way to early to make the call. It's difficult for me to believe Intel would blow 2 launches in a row.

It would be unlikely for them to not make any effort to address the uncore/fabric issues that Arrow Lake has as well, even if it's largely just increasing fabric clocks everywhere.

That is what I think as well. If all they do is fix the fabric issues (and get rid of some of the awful latency issues through a combination of CPU and fabric changes), I suspect performance will jump quite a bit.

Intel's been doing 2 gens per socket for so long. OEMs don't seem to mind.

Not sure why you think such a thing? You think OEM's like re-designs? Engineering costs lots of money. I'll bet OEM's like a drop in replacement even more than DIY does!

As an aside, shouldn't we be seeing some leaked Panther Lake benchmarks soon if consumers are to have them in their hands by the end of this year?

 

[511]

As an aside, shouldn't we be seeing some leaked Panther Lake benchmarks soon if consumers are to have them in their hands by the end of this year?

Yes the question is who is going to leak the stuff 🤣 cause I can't wait Strix Halo and PTL are the only two interesting products this Year(For me 18A is the most interesting one but it is a node not a product in the end it will be limited by your design).

 

[OneEng2]

Yes the question is who is going to leak the stuff 🤣 cause I can't wait Strix Halo and PTL are the only two interesting products this Year(For me 18A is the most interesting one but it is a node not a product in the end it will be limited by your design).

PTL for me.

I am very eager to see how 18A performs.

 

[AMDK11]

Question for 100 points. Is PantherLake a new core or slightly improved or the same as LionCove? Is NovaLake a different core than the one in Panther? Good question.