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+4+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.

 

[poke01]

Intel's Performance Index:

https://edc.intel.com/content/www/us/en/products/performance/benchmarks/intel-core-ultra-processors-series-3_1/

SKUs:

View attachment 136238

View attachment 136240

What I found interesting was for the M5 comparison, Intel set the PL1 to 65 watts for Panther Lake. For AMD, Intel and Qualcomm it was to set to 45W.



oh and single core power consumption got halfed. Its 10-12 watts now, before it was 20 watts.

 

[jdubs03]

Some interesting claims in there. If validated, it’s a good step. Though the X2E(E) will still be out of reach for now.

Up to 50% lower processor power with similar performance vs. AMD Strix Point with an Intel® Core™ Ultra X9 388H vs. AMD Ryzen™ AI 9 HX 370

As measured by Cinebench 2024 Single Core (PTL at 10W, AMD at 20W)​

Same power with similar performance vs. Qualcomm Snapdragon X Elite with an Intel® Core™ Ultra X9 388H vs. Qualcomm Snapdragon® X Elite X1E-84-100

As measured by Cinebench 2024 Single Core (at 20W).​

On par multithread performance vs. AMD Strix Point with an Intel® Core™ Ultra X9 388H vs. AMD Ryzen™ AI 9 HX 370

As measured by Cinebench 2024 Multi Core & Cinebench 2026 Multiple Threads (AMD @53W, Intel @ 45W) on par = +/-3%. .​

Up to 18% better multithread performance vs. Qualcom Snapdragon X Elite with an Intel® Core™ Ultra X9 388H vs. Qualcomm Snapdragon® X Elite X1E-84-100

As measured by Cinebench 2024 Multi Core & Cinebench 2026 Multiple Threads (Qualcomm at ~50W, Intel @ 45W).​

 

[mvprod123]

What I found interesting was for the M5 comparison, Intel set the PL1 to 65 watts for Panther Lake. For AMD, Intel and Qualcomm it was to set to 45W.

View attachment 136242

oh and single core power consumption got halfed. Its 10-12 watts now, before it was 20 watts.

It is interesting to look at third-party testing. 65 watts for "thin and light laptops"....meh.

 

[DavidC1]

oh and single core power consumption got halfed. Its 10-12 watts now, before it was 20 watts.

This is likely a misunderstanding. They are setting the PTL CPU lower power to a point where it performs similar, not that it peaks at 10-12W. It's like saying Alderlake uses 65W, while Rocketlake uses 250W - they are just setting Alderlake to be same performance as Rocketlake.

The post below you shows this:

Same power with similar performance vs. Qualcomm Snapdragon X Elite with an Intel® Core™ Ultra X9 388H vs. Qualcomm Snapdragon® X Elite X1E-84-100
As measured by Cinebench 2024 Single Core (at 20W).

The GPU performance seems really good though. 76% over Lunarlake at peak performance, but 50% in Steel Nomad at same power.

 

[poke01]

This is a misunderstanding. They are setting the PTL CPU lower power to a point where it performs similar, not that it peaks at 10-12W.

The post below you shows this:

makes sense now

 

[ToTTenTranz]

77% gaming performance uplift over Lunar Lake is immense for a single generation.


As for PTL-H in handhelds:




MSI, Acer and Microsoft (Xbox Claw? Surface?) are preparing gaming handhelds with PTL- H, plus all the main ODMs (Compal, Pegatron, Foxconn) which are probably going to be sold by other brands.

AMD has zero to counter this until MDS Premium, unless that 8c STX Halo is coming down in price pretty hard.

 

[jdubs03]

What I found interesting was for the M5 comparison, Intel set the PL1 to 65 watts for Panther Lake. For AMD, Intel and Qualcomm it was to set to 45W.

View attachment 136242

And still way over the 43W for the M5 in CB 2024.

I’m looking forward to the slide deck though. It’ll give us more context for the curves.

Edit: come think of it though based on:

On par multithread performance vs. AMD Strix Point with an Intel® Core™ Ultra X9 388H vs. AMD Ryzen™ AI 9 HX 370

As measured by Cinebench 2024 Multi Core & Cinebench 2026 Multiple Threads (AMD @53W, Intel @ 45W) on par = +/-3%.​

Depending on how you cut it, in CB 2024 the HX 370 gets around 1100-1130* in MT at 45-54W. The M5, around 1170. So it might be quite close there, then again the M5 has two less cores, even 2 more E-cores would’ve helped out a decent bit.

 

[poke01]

And still way over the 43W for the M5 in CB 2024.

it should be on par with AMD HX 370 and M5 at 45W PL1

 

[DavidC1]

Core Ultra 5 338H with ARC B370 and 10 Xe cores probably gets within 10% of the B390 chip while costing hundreds of dollars less for the laptop.

@ToTTenTranz That's at peak. Same power it does 50% better at Steel Nomad, which is still very good.

 

[mikk]

And still way over the 43W for the M5 in CB 2024.

I’m looking forward to the slide deck though. It’ll give us more context for the curves.

Edit: come think of it though based on:

Depending on how you cut it, the HX 370 gets around 1030 in MT at 45W. The M5, around 1170.

Pressdeck is here: https://download.intel.com/newsroom/2026/CES2026/Intel-CES2026-Core-Ultra-Series-3-Press-Deck.pdf

 

[regen1]

U5 338H should have pretty decent iGPU performance but it's been limited to LPDDR5X 8533MT/s(from the slides).

 

[DavidC1]

U5 338H should have pretty decent iGPU performance but it's been limited to LPDDR5X 8533MT/s(from the slides).

12% difference in bandwidth isn't significant, it'll be 5% at most. It's still going to be noticeably faster than HX 370 and Lunarlake, upwards of 30%. There was a leak of Pantherlake GPU, and 10% bandwidth only made 2% or so difference. Standard Intel iGPUs have never been seriously bandwidth bound. If you have 100% scaling with any one factor, then your design sucks(ie AMD iGPUs that "scale" 100% with memory at some point)

 

[Tigerick]

Geez, Intel has launched WCL as well with 46TOPS NPU + 18 TOPS GPU:

 

[jpiniero]

77% gaming performance uplift over Lunar Lake is immense for a single generation.

They're including Frame Gen in that, aren't they.

Only nVidia can do that

 

[DavidC1]

I'm going to say it again. That 4 Xe3 cores are going to match 8 Xe2 in many scenarios. Moore's Law and silicon advancements always benefit the lower end, lower power, and cheaper computers more than on the high end.

They're including Frame Gen in that, aren't they.

Only nVidia can do that

Yes, but the Lunarlake comparison also has frame gen. That said, 77% is 45W over 25W. At ISO-power, they are saying it's 50%.

 

[jdubs03]

Single-core levels off much more at the high-end than the 288V.

Multi-core does looks pretty solid though.

And yes I know they’re not the greatest graphs but the numbers can be deduced.

And it doesn’t quite match their comparison to the HX 370 (clearly shows higher than “on par”).

 

[regen1]

Geez, Intel has launched WCL as well with 46TOPS NPU + 18 TOPS GPU:

View attachment 136248
View attachment 136249

Those are low-end PTL SKUs and not Wildcat Lake. WCL's NPU has much lower net INT 8 TOPS(much below the current 40TOPS(INT 8) "Copilot+" threshold).

 

[Jan Olšan]

oh and single core power consumption got halfed. Its 10-12 watts now, before it was 20 watts.

That would be great but I think what they talk about is that the core generally has better 1t performance, so they can reduce it's clock/voltage/power to a lower than stock turbo boost level and achieve the same performance than a competing chip - this is about CInebench 2024 where Intel is strong. Hence why they list both 10W and 12W power for the same CPU (388H). That does not mean that actual 1T program power consumptio will be like this, since you will run the processor without the artificial cap.

You can see it in Jdubs03's pics.

Single-core levels off much more at the high-end than the 288V.
View attachment 136251
Multi-core does looks pretty solid though.
View attachment 136252
And yes I know they’re not the greatest graphs but the numbers can be deduced.

And it doesn’t quite match their comparison to the HX 370 (clearly shows higher than “on par”).

 

[poke01]

That would be great but I think what they talk about is that the core generally has better 1t performance, so they can reduce it's clock/voltage/power to a lower than stock turbo boost level and achieve the same performance than a competing chip - this is about CInebench 2024 where Intel is strong. Hence why they list both 10W and 12W power for the same CPU (388H). That does not mean that actual 1T program power consumptio will be like this, since you will run the processor without the artificial cap.

You can see it in Jdubs03's pics.

Thanks for the detailed reply. So we still need to wait for power figures for 1t at 5.1GHz.

 

[Tigerick]

Those are low-end PTL SKUs and not Wildcat Lake. WCL's NPU has much lower net INT 8 TOPS(much below the current 40TOPS(INT 8) "Copilot+" threshold).

It is. You should check the picture I posted. This is 25W WCL die; Intel should release low-power WCL later. At least Intel able to fit in NPU5 in the same SoC die (CPU+GPU+NPU)

 

[regen1]

They're including Frame Gen in that, aren't they.

Only nVidia can do that

77% claim vs U9 288V is on "2X Upscaling" but does it also have Frame-Gen in the mix ? Can't find the mention of Frame-gen in that.
Slides also claim 73% faster on 2X Upscaling and 82% faster on native rendering vs HX370.
Anyway we should wait for third party reviews for performance estimations.

 

[Hulk]

I refuse to spend 1 iota of brainpower analyzing Intel's BS graphs.

My Asus ProArtPX13 (HX370) scores 4945 in Cinebench 2026 MT while pulling 55W. I'll be curious to see if Panther Lake can hit that score at half the power.

 

[jdubs03]

I refuse to spend 1 iota of brainpower analyzing Intel's BS graphs.

My Asus ProArtPX13 (HX370) scores 4945 in Cinebench 2026 MT while pulling 55W. I'll be curious to see if Panther Lake can hit that score at half the power.

That’s not what they claimed. They claimed the same level of performance at 15% less power.

But either way we won’t have much time to wait for these claims to be debunked or verified, or somewhere in between.

 

[511]

Comparison against RTX 4050

 

[511]

That’s not what they claimed. They claimed the same level of performance at 15% less power.

But either way we won’t have much time to wait for these claims to be debunked or verified, or somewhere in between.

Need David Huang testing tbh I don't trust other reviewer with the details