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.

 

[DavidC1]

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

No you save money buying not being forced to buy Ultra 7 or Ultra 9 CPUs when wanting the top GPU config. Cause for the people who wants graphics performance the faster CPUs are just waste of money. It's like being forced to spend on a new platform rather than just swapping out the GPU in your desktop.

 

[adroc_thurston]

That speaks about your "sources".

It is true, PTL-H is slower than ARL-H.

No you save money buying not being forced to buy Ultra 7 or Ultra 9 CPUs when wanting the top GPU config.

No one wants the top iGP config because no one wants iGPs.

 

[DavidC1]

No one wants the top iGP config because no one wants iGPs.

Speak for yourself. And AMD has a dedicated line with their G series, but they still do the beancounter pairing peak GPU with 8700G rather than having it with 8300G.

 

[adroc_thurston]

Speak for yourself

I speak for the market.

And AMD has a dedicated line with their G series, but they still do the beancounter pairing peak GPU with 8700G rather than having it with 8300G.

It's not a dedicated line, they're just normal APUs on a desktop socket binned in a normal way.

 

[DavidC1]

It is true, PTL-H is slower than ARL-H.

This from the Zen 5 = 40% guy.

 

[511]

PTL will certainly match or beat ARL in ST perf without regression jmho(0-5%)

 

[adroc_thurston]

This from the Zen 5 = 40% guy.

Yeah, Turin's actually faster than that. too bad!

PTL will certainly match or beat ARL in ST perf without regression jmho(0-5%)

match, probably. beat? not with that fmax deficiency.
Doesn't help that the L3 is actually smaller gen on gen.

 

[511]

Yeah, Turin's actually faster than that. too bad!

not per core... and outside of some AVX-512 scenarios

 

[adroc_thurston]

not per core...

Yeah it is, F SKUs are a thing.
96c parts is also nuts (it's also what a peciluar CSP uses).

 

[LightningZ71]

People on here acting like 1T performance doesn't matter for games...

An i3 binned processor is going to be down almost a full Ghz if not more from the i9 bin for the best core. In the necessarily lower resolutions that iGPU gaming will require, that 1T performance will still be an issue, or do you REALLY want bad 1% lows?

 

[Fjodor2001]

Yeah, Turin's actually faster than that. too bad!

It was a general Zen5 claim that turned out to be incorrect.

 

[Fjodor2001]

No one wants the top iGP config because no one wants iGPs.

Nobody wants dGPU on mobile if iGPU performs close or has better perf/$, and you know it.

It's an AMD vs Intel case. AMD used to be iGPU king but not any longer. Same on DT.

 

[LightningZ71]

That was AMD's choice. They could have had better, but went a different direction.

 

[adroc_thurston]

Nobody wants dGPU on mobile if iGPU performs close or has better perf/$, and you know it.

no because larger iGP is a cost adder that no one wants.

 

[dullard]

People don't use iGPUs

That is just a lie and you know it.

 

[Fjodor2001]

That was AMD's choice. They could have had better, but went a different direction.

Agree. Bad move.

 

[Fjodor2001]

no because larger iGP is a cost adder that no one wants.

Using iGPU will be cheaper than adding corresponding separate dGPU. The customer can select whatever level of iGPU perf they want. E.g. Panther Lake will provide anything from 4 to 12 Xe3 cores. Everyone will be happy.

 

[jpiniero]

no because larger iGP is a cost adder that no one wants.

Oh they do. But they don't want to pay for it.

 

[Fjodor2001]

Oh they do. But they don't want to pay for it.

Well, they have the freedom to select whatever level of iGPU performance they want. Anything from 4 to 12 Xe3 cores. Whatever choice they make, it'll be cheaper than corresponding dGPU.

 

[Josh128]

Valley has been obsolete benchmark like forever now.

Not for mobile iGPU it isnt. I just ran it 3 days ago on a 12CU Ryzen 7 6800H in a Beelink NUC Win 11 PC and at max power it struggled to hit 30fps at 1080p low or medium with no AA. I was actually expecting much more from Zen 3+ / 12CU RDNA 2 chip but it very much struggled and dropped into mid to low 20s quite often.

 

[Tup3x]

Not for mobile iGPU it isnt. I just ran it 3 days ago on a 12CU Ryzen 7 6800H in a Beelink NUC Win 11 PC and at max power it struggled to hit 30fps at 1080p low or medium with no AA. I was actually expecting much more from Zen 3+ / 12CU RDNA 2 chip but it very much struggled and dropped into mid to low 20s quite often.

I would be surprised if any modern GPU has any optimisations for it. It's obsolete piece of software that hardly tells anything about real world gaming performance.

 

[Josh128]

I would be surprised if any modern GPU has any optimisations for it. It's obsolete piece of software that hardly tells anything about real world gaming performance.

?? My 6700XT that is older than the 6800H and scores over 12x the fps using the same settings. Optimizations have nothing to do with it, 6700XT and 6800H are same GPU architecture, and the 6800H is newer.

 

[Khato]

OEMs like options that increase their ASPs without requiring additional investment. Discrete GPUs fall into this category because they can design a laptop to accept a standardized discrete GPU card and then sell that laptop with iGPU only or 3-4 levels of dGPU card. Slotted memory and SSDs fall into the same category - no additional costs, just pure ASP increase for the upgrades. Such is much of the reason why the OEMs only begrudgingly offered LNL - it cut out one of their favorite money makers.

Now with respect to PTL iGPU options, OEMs like it because it doesn't really intersect with their dGPU offerings. Instead it's allowing them an additional configuration option in all the non-dGPU categories. You want your ultralight 13" to have better graphics? Pay an extra $50 to get the upgraded iGPU. With PTL Intel will let OEMs offer anything from 4+0+4+4Xe to 4+8+4+12Xe configurations in a single design and charge accordingly.

Note that the above is also related to the reason for the lack of widespread Strix Halo adoption. It's effectively 3 SKUs as the 380 is a bad joke. Going from the 385 to the 395 gets you 2x CPU cores and 1.25x GPU, aka who cares? Whereas going with a dGPU allows for easily spanning from an RTX 5050 to a 5090 in one design if desired, a 4x increase. Oddly I expect that AMD would have had better reception if they'd gone with on-package memory for Strix Halo to allow for more meaningful SKU differentiation and cheaper platform costs for the OEM (would have made the 380 a viable option.)

 

[adroc_thurston]

Whereas going with a dGPU allows for easily spanning from an RTX 5050 to a 5090 in one design if desired

nope, you need 3 separate platforms to go from the 5050 to 5090.
I'm not even talking about cooling assembly differences.

Oddly I expect that AMD would have had better reception if they'd gone with on-package memory for Strix Halo to allow for more meaningful SKU differentiation and cheaper platform costs for the OEM (would have made the 380 a viable option.)

MoP is Bad.

 

[DavidC1]

Using iGPU will be cheaper than adding corresponding separate dGPU. The customer can select whatever level of iGPU perf they want. E.g. Panther Lake will provide anything from 4 to 12 Xe3 cores. Everyone will be happy.

It's not cheaper, you actually pay more(quite a bit more) to get the high iGPU because you are forced to pay the highest CPU. With dGPU they can pair it with Core Ultra 3 if they want. And even 12 Xe3 is going to perform like the lowest end dGPUs anyway, so why do you need Ultra 9 to get it?

Want to spread halo iGPU? Get a SKU out that doesn't waste money on high end CPU. Right now you are paying $250 CPU + $100 GPU. How about $100 CPU + $100 GPU?

AMD/Intel: "Pairs their GPU with 7 and 9 series GPUs

"Muh nobody ain't buying our halo graphics!"

Uhh, it's a $1500 system for x60 graphics.....

Not for mobile iGPU it isnt. I just ran it 3 days ago on a 12CU Ryzen 7 6800H in a Beelink NUC Win 11 PC and at max power it struggled to hit 30fps at 1080p low or medium with no AA. I was actually expecting much more from Zen 3+ / 12CU RDNA 2 chip but it very much struggled and dropped into mid to low 20s quite often.

It doesn't matter that GTX 1050 Ti is not super duper faster than Lunarlake or Strix graphics. Especially in many newer games(where you need the frames), the iGPUs are often faster. That's the point, not that Heaven is relevant.

People on here acting like 1T performance doesn't matter for games...

An i3 binned processor is going to be down almost a full Ghz if not more from the i9 bin for the best core. In the necessarily lower resolutions that iGPU gaming will require, that 1T performance will still be an issue, or do you REALLY want bad 1% lows?

1GHz lower is almost nothing. One is running at 4GHz and the other is running at 5GHz, whoopdee doo. The biggest thing is the low CPU paired Halo will save you a bunch of money, which makes sense, rather than paying $600+ for x60 class performance. Of course the financially irresponsible can pair their RTX 5060 mobile with 9950X3D and 285K.