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

[Tigerick]

Wildcat Lake (WCL) Specs

Intel Wildcat Lake (WCL) is upcoming mobile SoC replacing Raptor Lake-U. 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 Q1 2026.

		Intel Raptor Lake U	Intel Wildcat Lake 15W?	Intel Lunar Lake		Intel Panther Lake 4+0+4
Launch Date		Q1-2024	Q2-2026	Q3-2024		Q1-2026
Model		Intel 150U	Intel Core 7	Core Ultra 7 268V		Core Ultra 7 365
Dies		2	2	2		3
Node		Intel 7 + ?	Intel 18-A + TSMC N6	TSMC N3B + N6		Intel 18-A + Intel 3 + TSMC N6
CPU		2 P-core + 8 E-cores	2 P-core + 4 LP E-cores	4 P-core + 4 LP E-cores		4 P-core + 4 LP E-cores
Threads		12	6	8		8
Max Clock		5.4 GHz	?	5 GHz		4.8 GHz
L3 Cache		12 MB		12 MB		12 MB
TDP		15 - 55 W	15 W ?	17 - 37 W		25 - 55 W
Memory		128-bit LPDDR5-5200	64-bit LPDDR5	128-bit LPDDR5x-8533		128-bit LPDDR5x-7467
Size		96 GB		32 GB		128 GB
Bandwidth				136 GB/s
GPU		Intel Graphics	Intel Graphics	Arc 140V		Intel Graphics
RT		No	No	YES		YES
EU / Xe		96 EU	2 Xe	8 Xe		4 Xe
Max Clock		1.3 GHz	?	2 GHz		2.5 GHz
NPU		GNA 3.0	18 TOPS	48 TOPS		49 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.

 

[branch_suggestion]

Seems peak ST performance will be neck and neck for Zen 5 and Lion cove at the supposed frequencies?

Yes, which is why AMD cheaped out on area again and settled for another balanced PPA core, since AMD knows they have X3D to fall back on and Turin is still gonna be comfortably ahead in server, plus Strix has the better ancillary IP vs ARL-H.
Volumemaxxing is a risky game.

 

[H433x0n]

A few %% above GLC in specint is where Zen4 and bros dwell.

Dingus on PHX2 goes to 3.7 and that's a mobile part.
Please.

It's a few % above RPC, where Zen 4 is behind.

BTW have you seen SRF benchmarks? Looks like SRF is now the most efficient X86 server CPU in existence.

The power efficiency story of the Xeon 6700E series is a great one and allows Intel to regain competitiveness on that front and in a number of cases outperforming 4th Gen EPYC in performance-per-Watt for both Genoa(X) and Bergamo.

Pretty good event for Intel.

 

[adroc_thurston]

It's a few % above RPC, where Zen 4 is behind.

?
Zen4 is still ahead in sir 1-copy.

Looks like SRF is now the most efficient X86 server CPU in existence.

Define efficient.
Looks barely faster than EMR in most cases which is pretty kaput.

 

[invisible_city]

Client variant no.

You don’t know that.

 

[H433x0n]

?
Zen4 is still ahead in sir 1-copy.

Sauce?

 

[adroc_thurston]

Sauce?

Get an SPR-MCC, get a Genoa, bench it.

Also did you mean this? This isn't comparing against Bergamo, but a 64c(!!) Genoa.

 

[gdansk]

It's a few % above RPC, where Zen 4 is behind.

BTW have you seen SRF benchmarks? Looks like SRF is now the most efficient X86 server CPU in existence.



Pretty good event for Intel.

No 128 core Bergamo?

Edit: no it's there

 

[adroc_thurston]

No 128 core Bergamo?

They're not suicidal.

 

[H433x0n]

Get an SPR-MCC, get a Genoa, bench it.

I'm talking about client, Skymont at the moment is a client only core.

 

[poke01]

Intel got CAMM2 motherboards. Will be interesting to see how these effect benchmarks.

 

[adroc_thurston]

I'm talking about client, Skymont at the moment is a client only core.

Client is tricky because I don't have PHX2 on hand.
Ask cheese, maybe he'll do SPEC runs. Eventually.

 

[branch_suggestion]

BTW have you seen SRF benchmarks? Looks like SRF is now the most efficient X86 server CPU in existence.

9754 still slaps it and does have the PPW edge still.
Direct comp is Turin-Dense due to how late it is so...

 

[Ghostsonplanets]

Did anyone got a good, high resolution, photo of PTL I18A wafer?

That's the neat part. They basically won't.

RPL lives!

Now with SKT, RPL, MTL and ARL-U are non-starters. That's a big issue for Intel.

 

[Markfw]

Get an SPR-MCC, get a Genoa, bench it.
View attachment 100422
Also did you mean this? This isn't comparing against Bergamo, but a 64c(!!) Genoa.

That is a slow... Genoa also, the 9554 is what they should be using.

What we really need is a few 3rd party review sites to bench them and give results. I don't believe anything either company shows at a preview.

 

[poke01]

Now with SKT, RPL, MTL and ARL-U are non-starters. That's a big issue for Intel.

Bette to kill your own part yourself before someone else does

 

[Hitman928]

Sauce?

Honey mustard, please.

On a more serious note, you can check the Anandtech 13900k review. It has the 7950x slightly ahead of the 13900k in SPECint IPC.

 

[mikk]

From this slide LNL-MX iGPU seems to be easily faster than MTL-H iGPU at the same power.

Lion Cove might be only 10% faster per clock than Skymont. This is so bad from the big core given how much bigger it is.

 

[H433x0n]

Client is tricky because I don't have PHX2 on hand.
Ask cheese, maybe he'll do SPEC runs. Eventually.

no, I'm going off of the Phoronix review that I linked.

Edit: Looks like Intel 3 is a legit node. That's super impressive for what was considered a meme SKU a few days ago.

 

[adroc_thurston]

no, I'm going off of the Phoronix review that I linked.

Phoronix testing methodology is grade A mess. Don't (unless specific tests).

Now with SKT, RPL, MTL and ARL-U are non-starters. That's a big issue for Intel.

Not my problem.

That is a slow... Genoa also, the 9554 is what they should be using.

Oh it is.
Which is why they're comparing 128c/128t SRF against 64c/128t Genoa.
They're brave. Not suicidal.

 

[invisible_city]

Get an SPR-MCC, get a Genoa, bench it.
View attachment 100422
Also did you mean this? This isn't comparing against Bergamo, but a 64c(!!) Genoa.

It is comparing 128 threads vs 128 cores, which is the correct comparison for hyperscalers that count 1 thread as a vCPU.

 

[Saylick]

Yes, which is why AMD cheaped out on area again and settled for another balanced PPA core, since AMD knows they have X3D to fall back on and Turin is still gonna be comfortably ahead in server, plus Strix has the better ancillary IP vs ARL-H.
Volumemaxxing is a risky game.

I wouldn’t call it cheapening out, but it’s a very engineer-y move and not out of line with what Zen was originally named to do: bring a balanced approach to x86 cores. Using balanced cores (read: high PPA) lets them win in servers, where it matters the most. But I will admit I was really hoping they’d make an x86 version of ARM’s X cores instead of iterating on top of a wildly successful middle core.

 

[SpudLobby]

Seems like it was AI generated lol

I think they are just saying that SMT traditionally gives 30% more throughput and 20% better power efficiency (so the power increases a little for that 30% more throughput).

But here's the relevant slides:
View attachment 100417
View attachment 100418
View attachment 100419

Awesome. Let’s go Intel.

 

[gdansk]

Intel 3 is a legit node

And technically just-on-time according to this 2021 chart from Anandtech.
Maybe Pat did it.

 

[adroc_thurston]

which is the correct comparison for hyperscalers that count 1 thread as a vCPU.

Let me introduce you to my friend AMD EPYC™ Bergamo 9754S.

 

[gdansk]

It is comparing 128 threads vs 128 cores, which is the correct comparison for hyperscalers that count 1 thread as a vCPU.

I have to ask, however, if you're a hyperscaler and you can buy a 128c Bergamo, why isn't it in the comparison?
Seems negligent.