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.

 

[regen1]

https://twitter.com/x/status/2017317708262740381

the results are out

Now some cross-comparison from this(for X7 358H v U7 255H) would imply SPEC INT Rate per GHz gains a bit higher than 8%(~8.35) but has caveats of different GCC versions, Kernel versions, etc.
Geekerwan and Bibalaptop got slightly under 10% and 6.7% respectively for their reference system configurations of U9 X388H v U9 285H.

 

[adroc_thurston]

M5
L1D: 128KBx4
L1I: 192KBx4
L2 shared: 8MB per cluster.
L3S: 8MB? M4 was 8MB

oh no M5 is uh, different.
It has a private L2 now (at 1M, idk latency).

 

[fastandfurious6]

-

 

[511]

I was equalizing it from ARL iso core count (4x vs 4x), but this works too.
Also SLC is slow as fk and pretty much every company that has a SLC (Qcomm + Intel) says it's not mainly about the CPU.
View attachment 137679

Uh the calculations are wrong no? It's around 16MB L2 for 2 Cores so for 4 won't it be 32.512 MB for Apple? and also we have SLC as well we are forgetting that in this equation Intel has 8MB SLC ARM vendors as well

 

[MoogleW]

Yea, based on the Novalake desktop leak of what being little over 10% faster than Arrowlake is not convincing.

Does that rumor still make sense?

Pantherlake IPC alone is one clue. Unless NovaLake has 2% IPC gain on a simultaneous tick and tock after achieving 5% on a tick. Especially on TSMC so there should not be a further regression in clock speeds

We also know NovaLake BLLC exists. Why would they bother if it makes no real difference in performance?

 

[poke01]

oh no M5 is uh, different.
It has a private L2 now

news to me. interesting, it hasn't been publicly reported.

You are missing E cores as well cause the 8E also have access to 18M L3 it's not for P core alone for 4 P core it would be 3 MB/core

i'm just focusing on P cores for now but yeah makes sense.

Uh the calculations are wrong no? It's around 16MB L2 for 2 Cores so for 4 won't it be 32.512 MB for Apple?

it would be 16MB L2 for 4 cores

 

[511]

it would be 16MB L2 for 4 cores

no? A18 Pro example

 

[Geddagod]

Uh the calculations are wrong no? It's around 16MB L2 for 2 Cores so for 4 won't it be 32.512 MB for Apple?

The Apple M4 has 16MB L2 for 4 cores. I think the problem with Apple's and Qcomm's config is that you have to have an acceptable amount of minimum L2.

and also we have SLC as well we are forgetting that in this equation Intel has 8MB SLC ARM vendors as well

Multiple vendors have stated that the SLC doesn't really exist for the CPU side. Doesn't matter too much.

 

[511]

The Apple M4 has 16MB L2 for 4 cores. I think the problem with Apple's and Qcomm's config is that you have to have an acceptable amount of minimum L2.

looks like it

Multiple vendors have stated that the SLC doesn't really exist for the CPU side. Doesn't matter too much.

are we sure though cause on Apple platform i know who to ask though

 

[Geddagod]

are we sure though cause on Apple platform i know who to ask though

I'm sure it can be accessed, just like it can be accessed for Qcomm's CPUs, it just doesn't impact perf much.
I mean, the X elite's SLC is only 6MB for example too, and they use a very similar caching hierarchy as Apple.

 

[511]

I'm sure it can be accessed, just like it can be accessed for Qcomm's CPUs, it just doesn't impact perf much.
I mean, the X elite's SLC is only 6MB for example too, and they use a very similar caching hierarchy as Apple.

Same Architect 🤣 🤣 also it is indeed 16MB l2 for 4 Cores very weird that it remains the same regardless of 2/4 Cores

 

[hemedans]

Really? So I guess you didn’t buy any Coffelake, Ice lake, Tiger Lake, Arrow Lake, Raptor lake, Meteor Lake laptops right?

Cause those above architectures sucked at battery life, temps etc and were horrible to use in laptops.

Tiger Lake and below those U and Y tier CPU they had good battery life, you had to choose though either low perfomance and good battery life or more perfomance and less battery life.

Nowadays modern Cpu offer both, you can play AAA games while using 15W or less.

 

[511]

Credit

https://twitter.com/x/status/2017528456058155265

 

[Fjodor2001]

Hvabehar

So what does all the non-English stuff say?

 

[511]

So what does all the non-English stuff say?

P-Core Comparison: UX9-388H vs. U7-258V​

SPECint Power Consumption Comparison at the Same FrequencyBoth have a maximum frequency of 4.8 GHz.

• 388H Group: Average IA (Instruction Architecture) power consumption for SPECint is 9.18W.
• 258V Group: Average power consumption for SPECint is 10.24W.

Performance and Efficiency:At 4.8 GHz, the performance (Perf) of the 388H is approximately 104% of the latter. Calculated out, the efficiency of the former is roughly 16% higher than the latter.

Translation from gemini

 

[Fjodor2001]

Perhaps they are talking about efficiency at different frequency, which accounts for the 16% vs 25%. Unclear what frequency was referred to for the 25%.

 

[511]

both the SKU have a max frequency of 4.8Ghz

 

[Fjodor2001]

So is it performing as expected, or better or worse?

 

[mvprod123]

https://twitter.com/x/status/2017636961650176156

 

[511]

8-10% Ipc is more than enough for yearly gain

 

[OneEng2]

The average people can also be quite ignorant. There needs to be more of a middle ground. OLEDs are ticking time bomb consumables with their burn in.

I also realized how vapid we can be. My Dell XPS 12 screen was BEAUTIFUL. It all lasted 1 month. After that I didn't think it was anything special at all. While you will notice your battery life, and the fan noise.

We have had this since Nov 2025 and no burn in so far. Seems like its safe for a while. I might just get on it and hide the task bar though.... just in case.

 

[Fjodor2001]

8-10% Ipc is more than enough for yearly gain

Yeah, at least it’s better than Zen4 -> Zen5%

Sure it was perhaps 7-8% in reality instead of 5%, but it was also more than 1 year between Zen4 and Zen5.

 

[511]

The biggest pro

Yeah, at least it’s better than Zen4 -> Zen5%

Sure it was perhaps 7-8% in reality instead of 5%, but it was also more than 1 year between Zen4 and Zen5.

Zen 4 -> 5 was like 16% IPC improvements but this is still decent

 

[Fjodor2001]

The biggest pro

Zen 4 -> 5 was like 16% IPC improvements but this is still decent

Right. Mixed up IPC with the gaming perf improvements for Zen5.

Anyway, Zen4->Zen5 was almost 2 years. So 16% in 2 years is about the same or actually less per year, since PTL got 8-10% in one year.

 

[511]

Right. Mixed up IPC with the gaming perf improvements for Zen5.

Anyway, Zen4->Zen5 was almost 2 years. So 16% in 2 years is about the same or actually less than 8-10% in one year which PTL got.

NVL E cores are getting 10%+ improvements as well