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.

 

[igor_kavinski]

The top NVL-S SKU is supposed to outnumber Zen 6 in core count by 2:1. I don't see why Cinebench would be a loss or even a narrow win for NVL.

It will be a new low for Intel if they claim a win at something like 300W or more.

 

[coercitiv]

It will be a new low for Intel if they claim a win at something like 300W or more.

No, it won't. That's 150W per tile, 5-6W per core. I would not be surprised if AMD also increases TDP for their top SKU, as long as the socket allows.

 

[igor_kavinski]

I would not be surprised if AMD also increases TDP for their top SKU, as long as the socket allows.

Still doubt that it would approach 300W.

 

[vanplayer]

I am having the same vibe as Zen5 vs ARL. Zen 6 is getting Hyped but NVL is meh in the end it will be similarish just like Zen5 vs ARL played out similar ST/MT.

Internal estimation is NVL loses everything(gaming, ST, etc) except MT against Zen6.

I'm actually expecting something good from the bLLC part.

bLLC doesn't exist. I'm afraid.

 

[511]

Internal estimation is NVL loses everything(gaming, ST, etc) except MT against Zen6.

are we refering that >1.1x slide?

 

[Geddagod]

are we refering that >1.1x slide?

The one that is talking about NVL?

 

[igor_kavinski]

bLLC doesn't exist. I'm afraid.

If that's true, they are screwed even more.

I just don't see how it cannot exist because they HAVE seen how the world went crazy for 9800X3D. Can't they hurriedly increase the cache size since they have experience with such large caches on their server SKUs?

 

[511]

The one that is talking about NVL?

Yes that is NVL

 

[511]

bLLC doesn't exist. I'm afraid.

it exists

 

[igor_kavinski]

it exists

OK, now you need to prove it because you sound so sure.

With real info. Not leaked.

(Send me the NDA'ed slide you got. I won't tell anyone I got it from you)

 

[vanplayer]

are we refering that >1.1x slide?

no

it exists

If that's true, they are screwed even more.

I just don't see how it cannot exist because they HAVE seen how the world went crazy for 9800X3D. Can't they hurriedly increase the cache size since they have experience with such large caches on their server SKUs?

well, yeah. I can't provide the source. These wouldn't be reported on any website because source is incognito.

Hint: NVL bLLC's non-existence is just like the SMT's absence in ARL, they're not seen in any Intel's documentation like software optimization guide.

 

[511]

Hint: bLLC's non-existence is just like the SMT's absence in ARL, they're not on any Intel's documentation like software optimization guide.

Well get better documents

 

[igor_kavinski]

Why would bLLC need to be documented? It's supposed to be transparent in operation.

 

[511]

Why would bLLC need to be documented? It's supposed to be transparent in operation.

Intel hasn't shared bLLC docs yet with customers

 

[igor_kavinski]

Intel hasn't shared bLLC docs yet with customers

Again, why would it need to be documented? You are not supposed to access caches directly via instructions.

This is not HBM cache.

Unless...

it is?

 

[511]

Again, why would it need to be documented? You are not supposed to access caches directly via instructions.

This is not HBM cache.

Unless...

it is?

bruh i meant Product documentation you went too deep into the cache now you are missing the point what a cache miss.

 

[Fjodor2001]

Internal estimation is NVL loses everything(gaming, ST, etc) except MT against Zen6.

are we refering that >1.1x slide?

no

So what source are you referring to?

 

[Fjodor2001]

Since NVL-S outnumbers Zen6 by 2:1 in core count for the top SKU, it should be possible to run the former at a bit lower frequency when maxing MT perf and still win in MT.

Squeezing out the last bit of perf by pushing towards max frequency consumes a lot of power. Just dropping the frequency a bit from max reduces power consumption a lot, while not losing much performance. So NVL-S could win in MT while still not consuming huge amounts of power (in relation to the MT perf it delivers). But let’s see how it turns out.

 

[511]

So what source are you referring to?

I use the best source

Spoiler: Spoiler

Trust Me Bro

 

[LightningZ71]

You really must've paid a pretty penny for it. 1.4 GHz Tualatin wasn't exactly a widely available CPU.

At the time, I had a close friend working at a local white box computer shop that was able to get it at cost, but it did take some doing. That thing lasted me for a long time.

 

[gdansk]

I am having the same vibe as Zen5 vs ARL. Zen 6 is getting Hyped but NVL is meh in the end it will be similarish just like Zen5 vs ARL played out similar ST/MT.

Do you remember how that ended? People were so damn confident ARL had it after Zen 5% that they're still coping by comparing tweaked ARL to untweaked GR. Today Intel is selling 20 cores that cost much more to make than 1CCD GR for $210 here. You know it's "similarish" when retailers have to undercut the competition's part by ~$150 to move inventory.
I would not want a repeat of that.

And if NVL ends up similar to Zen 6 it's even worse than ARL, because this time they're throwing huge amounts of silicon at cache and multi-tile configurations.

 

[jpiniero]

Since NVL-S outnumbers Zen6 by 2:1 in core count for the top SKU, it should be possible to run the former at a bit lower frequency when maxing MT perf and still win in MT.

Cinebench accelerators tho.

 

[Fjodor2001]

Cinebench accelerators tho.

No, because we're talking about MT perf.

 

[reb0rn]

Cinebench accelerators tho.

Its not, you then do not need new CPU and never will, I do need more MT and some ppl that do real work and not just gaming!

this talking about we do not need more MT speed is like childish at best

 

[gdansk]

Its not, you then do not need new CPU and never will, I do need more MT and some ppl that do real work and not just gaming!

this talking about we do not need more MT speed is like childish at best

What real world, big boy work are you doing that doesn't need to communicate outside of its L2 ghettocluster quickly?
The current monts in desktop parts are configured for one thing and one thing only - chart wankery.