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.

 

[adroc_thurston]

More PCIe lanes.

not what I meant by that.

It's the same just one has stuff disabled cause to stop OEM from doing idiot stuff like pairing 12 Xe3 iGPU with dGPU.

ah okay. Crisis averted?

The main thing that sucks anyway Stephen Robinson is the chief architect now so I have hope

UC is bretty far away doe.
gonna be rough.

 

[511]

PTL is good overall they are sharing Panther GPU Tiles with Nova as well there is certainly lot of reuse.

 

[adroc_thurston]

they are sharing Panther GPU Tiles with Nova as well there is certainly lot of reuse.

wait what.
No Xe3P outside of -AX?

 

[DavidC1]

Finally using their brains they could have the same SoC tile for Desktop and Laptop ARL/MTL but they had to mess it up.
NVL tile arch is way better they ARL/MTL

Intel has smart people, it's just there's a lot of infighting and constant change of plans. And you add DEI-hiring policies on top... It's like in Warcraft 3 where you have limited amount of numbers and you charm one enemy unit to your side and suddenly the whole battle changes.

If there are 100 people, and there are two sides equally split between 50 people each, it's effectively like there's no one working cause they cancel each other out.

 

[adroc_thurston]

It's like in Warcraft 3 where you have limited amount of numbers and you charm one enemy unit to your side and suddenly the whole battle changes.

you definitely did not play WC3 on bnet at all.

 

[511]

wait what.
No Xe3P outside of -AX?

and 12 Core Xe3P yes NVL Desktop will be 4Xe3 which will still be big uplift vs 4Xe and it's fine for Desktop.

 

[ToTTenTranz]

Their own slides claim ~+50% over Lunar Lake. Lunar Lake is ~ Strix Point (being a little generous here) for real gaming fps. Strix Point is ~GTX 1050Ti for gaming. GTX 1080 is 2.5-3X faster than 1050Ti.

Lunar Lake is faster than Strix Point when it's not hampered by drivers, and Strix Point is a lot faster than the GTX 1050 Ti:








And this is using pure raster which, to be frank, few people are using on modern hardware in modern games.

 

[511]

https://twitter.com/x/status/1976942300355084339

 

[Fjodor2001]

For the U-range, what improvements are expected going from Panther Lake to Nova Lake? I wonder if it's worth waiting for the latter.

 

[511]

For the U-range, what improvements are expected going from Panther Lake to Nova Lake? I wonder if it's worth waiting for the latter.

for the cpu apx/avx 10.2

 

[regen1]

For the U-range, what improvements are expected going from Panther Lake to Nova Lake? I wonder if it's worth waiting for the latter.

For U-class/4P+0E+4LPE config
P- core changes to Coyote Cove, (E)/LPE- core changes to Arctic Wolf with APX, AVX10.2 support, CPU tile(4P+0E) on 18A-P node
New media engine
Integrated Thunderbolt 5
New NPU
Potential improved memory speed support

 

[Fjodor2001]

Thanks. Not that many improvements with Nova Lake after all then, for the U series. Might go with a Panther Lake based laptop then. But will be waiting for reviews first.

Seems like there are move improvements coming on the desktop side side for Nova Lake, with NVL-S.

 

[Josh128]

Lunar Lake is faster than Strix Point when it's not hampered by drivers, and Strix Point is a lot faster than the GTX 1050 Ti:




View attachment 131786
View attachment 131787


And this is using pure raster which, to be frank, few people are using on modern hardware in modern games.

These are cherry picked, and this is GTX 1050 Ti Mobile. Im talking real GTX 1050 Ti desktop. When not VRAM limited, its usually as fast or can be faster than 890M, as seen below.




But I digress, its hard to find decent comparisons with 1050Ti due to the age gap. GTX 1060 desktop is definitely faster than the 890M, and GTX 1080 is ~50%-75% faster than GTX 1060. Panther Lake will not be matching 1080 in real gaming scenarios.

890M vs 1060

1060 6GB vs 1080

W

 

[ToTTenTranz]

These are cherry picked, and this is GTX 1050 Ti Mobile. Im talking real GTX 1050 Ti desktop.

1050 Ti Mobile and desktop have practically the same performance.

When not VRAM limited,

It has 4GB VRAM. It's pretty much always VRAM limited in any modern game.

its usually as fast or can be faster than 890M, as seen below.

That's a synthetic benchmark. I showed actual videogames.

GTX 1060 desktop is definitely faster than the 890M, and GTX 1080 is ~50%-75% faster than GTX 1060.

The more recent games in that video, Forza Horizon 5 and Cyberpunk 2077, show better performance on the 890M. And that's before applying FSR3.

Panther Lake will not be matching 1080 in real gaming scenarios.

It will probably surpass the GTX1080 when using XeSS at lower resolutions.
The thing with desktop Pascal is it doesn't even do double-rate FP16 so the architecture is completely hopeless even for upscalers, let alone ML-based ones.

 

[Tup3x]

These are cherry picked, and this is GTX 1050 Ti Mobile. Im talking real GTX 1050 Ti desktop. When not VRAM limited, its usually as fast or can be faster than 890M, as seen below.


View attachment 131854

But I digress, its hard to find decent comparisons with 1050Ti due to the age gap. GTX 1060 desktop is definitely faster than the 890M, and GTX 1080 is ~50%-75% faster than GTX 1060. Panther Lake will not be matching 1080 in real gaming scenarios.

890M vs 1060

1060 6GB vs 1080

W

Valley has been obsolete benchmark like forever now.

 

[DavidC1]

These are cherry picked, and this is GTX 1050 Ti Mobile. Im talking real GTX 1050 Ti desktop. When not VRAM limited, its usually as fast or can be faster than 890M, as seen below.

I showed you the link comparing gaming performance against 1050 Ti Desktop. It's in average 10% faster. Newer games bigger advantages, with much as 30% gains. It's not my fault you chose to ignore it. Lunarlake is even faster with driver updates and the change to having PL2 at least 1W higher than PL1.

And Intel says Xe3 is >50% faster than Lunarlake. > means Greater than. It looks like it might be 70-80%. At that point it's practically 1080 levels. Even if 10-20% behind it won't matter because it's close enough, since I have one and it uses 170-190W in games to do while Pantherlake can do it at 35W including the CPU. Modern gaming performance is more relevant anyway since older games run fast enough already.

Only real issue with Xe3 is that I have to spend a grand american dollars whereas you can buy used 1080 for $120.

 

[LightningZ71]

The most relevant comparison points should be the mobile 1650 and the high memory 3050 mobiles. The next step should be the mobile 4060 8gb. Those were VERY high volume parts and are going to be either up for replacement or what most users currently have.

The 890m, when given enough power and cooling, tends to hang with the full spec 1650 mobile parts in most games, especially when implementing FSR/XeSS.

 

[DavidC1]

The most relevant comparison points should be the mobile 1650 and the high memory 3050 mobiles. The next step should be the mobile 4060 8gb. Those were VERY high volume parts and are going to be either up for replacement or what most users currently have.

The 890m, when given enough power and cooling, tends to hang with the full spec 1650 mobile parts in most games, especially when implementing FSR/XeSS.

And it'll be close enough to RTX 3050 laptop that most will choose Pantherlake.

They said they addressed the issue where too much CPU is allocated, a longstanding Intel problem that existed since Extreme Graphics days. I saw one video where Intel had the core running at 4GHz whereas the faster AMD was running at 1.3GHz for the CPU.

I think a 12 Xe3P Novalake Core Ultra 3 series would be great. But again, too much logic for beancounters.

 

[LightningZ71]

We used to somewhat fix that with throttle stop.

 

[DavidC1]

With Pantherlake it also has an interesting feature where the shader compilation is streamed in from their cloud server and then downloaded to a specific folder. It reduces load times and stuttering. Pantherlake is better than I expected.

 

[adroc_thurston]

I think a 12 Xe3P Novalake Core Ultra 3 series would be great

You mean like an i3 with the full gfx blob? Lmao.

 

[DavidC1]

You mean like an i3 with the full gfx blob? Lmao.

Precisely. What's funny about it?

 

[adroc_thurston]

What's funny about it?

People don't use iGPUs so forcing them into the full iGP config pricing at i3 is lol.

 

[DavidC1]

People don't use iGPUs so forcing them into the full iGP config pricing at i3 is lol.

Lolz.

You claimed this in the Novalake thread:

It is NVL-S.
Panther Lake has 0 to negative % of 1t uplift over ARL/LNL.

That speaks about your "sources".

 

[S'renne]

Lolz.

You claimed this in the Novalake thread:

That speaks about your "sources".

I didn't know fat iGPUs can suddenly be that much cheaper that putting it on i3s keeps it at i3 pricing