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

[Tigerick]

Wildcat Lake (WCL) Preliminary Specs

Intel Wildcat Lake (WCL) is upcoming mobile SoC replacing ADL-N. 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 Q2/Computex 2026. In case people don't remember AlderLake-N, I have created a table below to compare the detail specs of ADL-N and WCL. Just for fun, I am throwing LNL and upcoming Mediatek D9500 SoC.

		Intel Alder Lake - N	Intel Wildcat Lake	Intel Lunar Lake		Mediatek D9500
Launch Date		Q1-2023	Q2-2026 ?	Q3-2024		Q3-2025
Model		Intel N300	?	Core Ultra 7 268V		Dimensity 9500 5G
Dies		2	2	2		1
Node		Intel 7 + ?	Intel 18-A + TSMC N6	TSMC N3B + N6		TSMC N3P
CPU		8 E-cores	2 P-core + 4 LP E-cores	4 P-core + 4 LP E-cores		C1 1+3+4
Threads		8	6	8		8
Max Clock		3.8 GHz	?	5 GHz
L3 Cache		6 MB	?	12 MB
TDP		7 W	Fanless ?	17 W		Fanless
Memory		64-bit LPDDR5-4800	64-bit LPDDR5-6800 ?	128-bit LPDDR5X-8533		64-bit LPDDR5X-10667
Size		16 GB	?	32 GB		24 GB ?
Bandwidth			~ 55 GB/s	136 GB/s		85.6 GB/s
GPU		UHD Graphics		Arc 140V		G1 Ultra
EU / Xe		32 EU	2 Xe	8 Xe		12
Max Clock		1.25 GHz		2 GHz
NPU		NA	18 TOPS	48 TOPS		100 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.

 

[Magio]

ARL-H was 40K .. and this has 1.5X Cores and higher frequency

Still, you're not getting 50% increase in perf at equal thermal envelope just from going from N3B to N3E. And arch improvements are more likely to materialize in game performance than in synthetic benchmarks. So 25% from N3E + wider GPU seems reasonable, until we learn more about TDP and how performance scales across that line.

Benchleaks has shared a few more results since, with a 53000 best score. Hoping we get first CPU results soon.

 

[fastandfurious6]

panther GPU slightly better than 890M (strix point)

how will it fare on CPU and thermals?

 

[511]

Still, you're not getting 50% increase in perf at equal thermal envelope just from going from N3B to N3E. And arch improvements are more likely to materialize in game performance than in synthetic benchmarks. So 25% from N3E + wider GPU seems reasonable, until we learn more about TDP and how performance scales across that line.

Benchleaks has shared a few more results since, with a 53000 best score. Hoping we get first CPU results soon.

ARL iGPU is N4P

 

[Covfefe]

ARL iGPU is N4P

The 50% faster claim is Panther Lake vs Lunar Lake.

 

[dullard]

At 2.5 GHz, that is 5% to 30% higher frequency than Intel iGPUs usually run. I don't remember any mobile iGPU running faster than 2.35 GHz. The equivalent lower Ultra 7 Lunar Lake (258V) was 1.95 GHz.

There were actually 8 iGPU benchmarks done on the 358H. Here is the best vs. the worst:

ASUS ROG Zephyrus G14 GU405AA_00016437B vs ASUS ROG Zephyrus G14 GU405AA_00016437B - Geekbench

browser.geekbench.com

And here is a 358H benchmark vs the best 258V Lunar Lake score that I could quickly find in a roughly equivalent ASUS 14" laptop (67% higher for the 358H):

https://browser.geekbench.com/v6/compute/compare/5050036?baseline=4673858

 

[ToTTenTranz]

https://twitter.com/x/status/1980465970196976099

PTL GPU Benchmark Leak

16GB VRAM allocation? System RAM detects 32GB, meaning they're allocating half of it for the iGPU which is a lot.

 

[dullard]

16GB VRAM allocation? System RAM detects 32GB, meaning they're allocating half of it for the iGPU which is a lot.

That is how that benchmark worked on Lunar Lake too. Takes half the memory.

 

[eek2121]

Looking forward to some actual gaming benchmarks!

And my point is this is the behavior of value seeking users, not people who buy very expensive laptops and dual 5k monitor setups. I understand there's a use case for people who want just one device (due to personal choice or work conditions) but the overlap between these people and the "dual 5k monitors + M2 external storage" is slim to none IMHO. The value seekers are fine with USB-C, TB is nice to have for them but may only be available out of their comfort range.


And the external M2 storage via TB is more useful for business users?

Meh, I don’t know about multiple 5k displays, but my “ancient” Zen 3 laptop has no issues driving the Apple Studio display.

I don’t think I have ever worked at a place where they gave you “premium” monitors. Nearly all of them bought cheap 1080p monitors for everyone.

 

[regen1]

 

[coercitiv]

Bad dad joke at the start, good engineer confirmed.

 

[511]

He is the chief architect of Intel x86.
I have enjoyed all of Robinson's talk like all of them

 

[511]

The 50% faster claim is Panther Lake vs Lunar Lake.

View attachment 132342

https://edc.intel.com/content/www/us/en/products/performance/benchmarks/intel-technology-tour-itt-2025/

 

[dullard]

First Arrow Lake Refresh benchmark is now out.

https://videocardz.com/newz/intel-core-ultra-7-270k-plus-first-arrow-lake-refresh-sku-leaks-out

Looks to me like the Ultra 7 270K is the Ultra 7 265K with 4 more E cores and a slightly lower base clock. Videocardz calls it the Ultra 9 285K with a lower turbo clock. Either way, it is in between the 265K and the 285K based on specs.

Memory speed gets a boost from DDR5-6400 to DDR5-7200. That'll be relevant to the vast majority of computer users who don't upgrade CPUs on the same motherboard. The 270K Seems to perform MT a few percent better than the 285K in an almost equivalent system (memory speed changed).

LENOVO INVALID vs LENOVO INVALID - Geekbench

browser.geekbench.com

 

[DavidC1]

panther GPU slightly better than 890M (strix point)

how will it fare on CPU and thermals?

Lunarlake 140V is already better than 890M.

 

[adroc_thurston]

Lunarlake 140V is already better than 890M.

oh no, not in video games it ain't

 

[ToTTenTranz]

oh no, not in video games it ain't

The 140V is faster than the 890M in every game that was picked up by Intel's driver team. And many of these results should be outdated as Intel has been making strides in their drivers pretty regularly.

Computer Games on Laptop Graphics Cards

News, Reviews and other Informations about Laptops

www.notebookcheck.net

It's also interesting to see the 890M's anemic bandwidth at play, where it starts much faster than the 140V at "ultra low" settings but at high it becomes slower. The F1 results are impressive.

 

[Magio]

oh no, not in video games it ain't

It depends, it does better in some, worse in others and similar in many. 140V does usually do it at better efficiency which means that in similarly thermally limited conditions it's the better pick.

What PTL looks to do is make it a concrete win at LNL-level TDPs while having a lot more upside at higher draw which really should make it not a contest anymore in any conditions.

 

[DavidC1]

oh no, not in video games it ain't

MSI Claw 8 AI+ A2VM Review

The MSI Claw 8 is a high-end PC handheld equipped with beefy graphics, 1 TB of storage, 32 GB of fast memory, a massive 8-inch screen, and a huge battery. Its gaming prowess is impeccable, but is the software good enough to warrant a recommendation? Find out in our review.

www.techpowerup.com

Yes it does, not only that, it's even better in handhelds, meaning power constrained environments.

140V does usually do it at better efficiency which means that in similarly thermally limited conditions it's the better pick.

That's why ISO-ing always matters. Cause otherwise you can compare it with 600W RTX 5090.

It's also interesting to see the 890M's anemic bandwidth at play, where it starts much faster than the 140V at "ultra low" settings but at high it becomes slower. The F1 results are impressive.

It can't be bandwidth only, because in Time Spy, which represents very demanding low fps scenarios, the scores go from 62xx using 8533MT/s to 63xx using 9600MT/s. By that figure, Pantherlake can perform 30-50% higher than that before it even starts reaching the level equal to normal GPUs, nevermind bandwidth constrained ones. Intel's normal iGPUs were never really bandwidth constrained, unless you suddenly dropped to single channel, or 2x'ed it like Iris without eDRAM.

It could be other factors at play with F1 2025 as well. Weaknesses of ARC graphics does carry a bit into their iGPUs as well. It's typically weak in GTA V, and in NBC test, even 140V is.

 

[MS_AT]

Yes it does, not only that, it's even better in handhelds, meaning power constrained environments.

Not to nitpick, but can you show where in this techpowerup review comparisons are made against 890m? AFAIK both handhelds they are comparing against are using zen4 chips. It is also worth pointing out that competing devices have half the RAM.

 

[hemedans]

Not to nitpick, but can you show where in this techpowerup review comparisons are made against 890m? AFAIK both handhelds they are comparing against are using zen4 chips. It is also worth pointing out that competing devices have half the RAM.

Phawx also have plenty of benchmarks

 

[coercitiv]

It's also interesting to see the 890M's anemic bandwidth at play, where it starts much faster than the 140V at "ultra low" settings but at high it becomes slower. The F1 results are impressive.

Notebookcheck's data is difficult to compare directly. For the 890M they tested gaming performance even with NUCs that use DDR5 5600MT/s. Even when the mem bandwith is theoretically the same, results vary greatly from one device to another.

To be clear, I'm not arguing this would make the 890M more of a winner, just pointing out evaluating the 890M needs more effort there or a more consistent source.

 

[DavidC1]

Not to nitpick, but can you show where in this techpowerup review comparisons are made against 890m? AFAIK both handhelds they are comparing against are using zen4 chips. It is also worth pointing out that competing devices have half the RAM.

Intel Lunar Lake iGPU analysis - Arc Graphics 140V is faster and more efficient than Radeon 890M

An analysis by Notebookcheck of the Lunar Lake Core Ultra 7 258V coupled with the new Intel Arc Graphics 140V iGPU.

www.notebookcheck.net

Only way for Pantherlake to be slower than 890M would be if Intel regressed. Those are also based on TDP settings, where in the platform level, Lunarlake is lower power than Meteorlake and Strix. So far, if there are weaknesses, graphics doesn't seem to be one.

Also, doubling of the GPU cache means it has 16MB, which is same as Infinity Cache of 6500XT. Someone was talking about how 890M could have fit 16MB if it wasn't for the DWU? Well... it seems we got an Intel version.

 

[MS_AT]

Only way for Pantherlake to be slower than 890M would be if Intel regressed. Those are also based on TDP settings, where in the platform level, Lunarlake is lower power than Meteorlake and Strix. So far, if there are weaknesses, graphics doesn't seem to be one.

Also, doubling of the GPU cache means it has 16MB, which is same as Infinity Cache of 6500XT. Someone was talking about how 890M could have fit 16MB if it wasn't for the DWU? Well... it seems we got an Intel version.

I don't recall mentioning Panther Lake anywhere. I just asked where are the supposed comparisons between 890m and Lunar Lake GPU in the techpowerup review you linked. That's all. I think it's a very precise question, you can link to a page from the review where the comparison can be seen or you can say that the review does not compare them. I just wasn't able to find it myself, but I might have missed something.

 

[LightningZ71]

the 6500XT had 18Gbps ram at 64 bits giving 144GB/sec bandwidth.
If Panther Lake is configured with LPDDR 8666 ram at 128 bits, it'll have 138 GB/sec bandwidth. any faster and it'll match or surpass the 6500XT.

I dare say that it'll be quite a disappointment if the iGPU on Panther Lake isn't capable of putting up numbers that are in the immediate vicinity of the 6500 XT on most gaming benchmarks. Depending on the game, and excluding any raytracing, the 6500XT is anywhere from 10+% to 20+% faster than the 890m at 1080P High.

Personally, I expect Panther Lake to slot in a bit above mobile 6500 XT specs and performance, with the ability to enable a few of the les demanding ray tracing features. It should do even better when both have XeSS / FSR enabled.

 

[regen1]

From Q3 earnings call:

In Client, we are on track to launch our first Panther Lake SKU by year-end, followed by additional SKUs in the first half of next year.