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.

 

[ajsdkflsdjfio]

https://twitter.com/x/status/1863779690546147767

Something i agree with

Just look at the messaging from the board

"Bloomberg reported that directors were frustrated with Gelsinger’s slow progress in chasing Nvidia’s lead"

"As a board, we know first and foremost that we must put our product group at the center of all we do"

What they're saying makes very little sense in the wake of Pat's departure. Seems like they have no idea what to do except "do an Nvidia" which is impossible to do at this stage. Also to repeat for the nth time, I don't see how CCG and DC/AI's failures are being put solely on Pat. Even the vast majority of products that released in 2024 were largely defined before Pat stepped in. Sure Pat was CEO for the validation and finalization of the products but he can't magically fix designs that were in place since before he was even there on top of fixing foundry. Let's be honest the product groups aren't performing worse than they have under the previous three Ceos arguably in fact they have been performing better recently as they have managed to release their products only a quarter or two late. It's just these investor slave directors can only tell when the profits are down not their momentum. It's like a driver breaking a car based only on it's position and not factoring it's velocity or its acceleration. Many of these same directors were present when intel was still "on top" while they were floundering on 14nm and 10nm for almost a decade, 2 nodes 10 years. They were also present while intel were divesting their billions of dollars into useless businesses while simultaneously ignoring their core business with client and datacenter missing their roadmaps by years rather than months.

Note how the foundry roadmap is still accurate, perhaps for the first time in the millennium, with 7++ (intel 3) in 2023 and 5++ (intel 18a) in 2025. Then the board of directors decide to remove the only CEO who has ever brought a division of intel back to working order.

 

[511]

Their roadmap is fine now so they ain't dying. But again, this is the wrong move.

That's Pats roadmap but thay ain't got anything after it 😂

 

[Gideon]

https://twitter.com/x/status/1863779690546147767

Something i agree with

Yeah ok, the board much worse than i thought. From the comments on the tweet

Besides the Boeing CFO/COO there’s the ADSK chair (most maligned software co with poor oversight of CEO). The M&A guy also let PYPL’s Schulman run PYPL kinda into the ground fwiw

And regarding NVIDIA in comparison:

I have an entire chapter in THE NVIDIA WAY on the need for technical engineering acumen over business MBA finance types. The consulting suits are a bad disease. The incredible thing is I interviewed dozens of former Nvidia employees and NONE of them had business suits backgrounds

 

[fastandfurious6]

they're not dead but this really is the wrong move.

beginning of the end, the corrupted in power literally said "we're going back to the way we were making short term profits to the detriment of any future"

They won't have enough time to rebound ever again

also market conditions are WAY more aggressive than Ever - basically if AMD Bulldozer happened in 2024 instead of 10 years ago etc, AMD would be finished no matter what.

 

[adroc_thurston]

the corrupted in power literally said "we're going back to the way we were making short term profits to the detriment of any future"

Well the whole point is that they're not making any profits to begin with.

 

[Gideon]

Well the whole point is that they're not making any profits to begin with.

Not a problem. Load it with additional (government) debt, claiming to be the only hope for western fabs. And in the background divest and sell off everything, close campuses retire "useless expensive opinionated engineers" (the Boeing CFO/COO should know) . I bet those EUV machines would make quite a dime from TSMC. And do a quick stock dump and pump while you're at it. Should fetch quite a lot (for the guys driving it to the ground) before bankrupcy.

Obviously this needs to coupled with obsfurcations, vehement denials and loads of fake news regarding future products and nodes, riding on the high of actual products that will still come out regardless e.g. Battlemage, Clearwater Forest, etc ...

 

[SmokSmog]

 

[poke01]

Jensen is a notorious penny pincher. There's no other explanation for his obsession with 8GB. The idiot even re-released 3060 with 8GB despite originally launching that GPU with 12GB.

Jensen's anything but that. He's a shrewed businessman, simply put. Nearly all his hires have been engineers, complete opposite of Intel.
If Jensen was a penny pincher, he won't spend 700+mm^2 dies on consumer GPUs. The 8GB cards are simply Nvidia taking advantage of their market power, which Intel should put a stop to those cards with BMG.

 

[misuspita]

I said this and I will say it a million times. These are the 💩heads that for a decade boughtback shares instead of investing in the company future and investing in the primary money maker, CPU. Yeah, missing AI is on Gelsinger but now it's too late. Even AMD who is better prepared and already having products can barely compete with Nvidia. You can say a lot about Jensen but he is a visionary. His cousin is also damn good. But what's happening at Intel is just the last tremors of a dying corpse, if they don't hire someone to straighten the ship.

They remind me of some failed games I play, where I try to invest a lot and at the end of the day I fail because my investments don't yet make enough money to pay for the debt I cause... Their fabs are expensive to maintain even if sitting idle.

 

[Hulk]

5 CEO's in 4 years. Intel is on track for that at least.

 

[AcrosTinus]

I said this and I will say it a million times. These are the 💩heads that for a decade boughtback shares instead of investing in the company future and investing in the primary money maker, CPU. Yeah, missing AI is on Gelsinger but now it's too late. Even AMD who is better prepared and already having products can barely compete with Nvidia. You can say a lot about Jensen but he is a visionary. His cousin is also damn good. But what's happening at Intel is just the last tremors of a dying corpse, if they don't hire someone to straighten the ship.

They remind me of some failed games I play, where I try to invest a lot and at the end of the day I fail because my investments don't yet make enough money to pay for the debt I cause... Their fabs are expensive to maintain even if sitting idle.

By that logic AMD would have been dead years ago.

Desktop is down but only in DIY, performance cannot reach AMDs x3D but for anyone on a 4070 or less which is nearly 70% it does not matter, system integrators will run through their raptor lakes and start using Arrow Lake, the average consumer will not give a damn. Furthermore I see no chance of Intel catching up to AMD in gaming without a break through, that battle is lost for maybe until the end of the decade.

On mobile, they are showing signs of live. The products are competitive and the OEMs are still giving Intel the better designs. In my company for example all notebooks are on Intel and that will stay that way, even legacy bare mental is still on Intel with potential shift to Epyc but I will veto that .

On AI, Gaudi 2 seems to focus on inference which should be the value engine in the future, the benchmarks show competitive performance as well as One API with SYCL.

On HEDT, it is over, Threadripper is just better... Maybe Intel will try again, RDIMMS are not nice for consumer budgets.

On Server, again signs of live, but against Epyc they only win in memory throughput and latency otherwise is it a bloodbath.

Overall not good but also not as bad as people make it out to be. Should 18A and 18AP really be as good or better than TSMCs equivalent and the designs be up to par, Intel would be back in the race.

 

[cannedlake240]

Intel still thinking of themselves as a chip design powerhouse they used be back in the day lol. Their attempts at GPUs and CPU cores clearly demonstrate they're completely washed up. Trying to suddenly pivot to products from foundry right at this moment seems like a mistake.

 

[oak8292]

Intel still thinking of themselves as a chip design powerhouse they used be back in the day lol. Their attempts at GPUs and CPU cores clearly demonstrate they're completely washed up. Trying to suddenly pivot to products from foundry right at this moment seems like a mistake.

Intel is foundry first. Robert Caldwell a retired Intel architect made a presentation in 2013 at Hot Chips about the focus at Intel. Architects were third after Moore’s Law (Dennard’s scaling) and marketing.

https://old.hotchips.org/wp-content/uploads/hc_archives/hc25/HC25.15-keynote1-Chipdesign-epub/HC25.26.190-Keynote1-ChipDesignGame-Colwell-DARPA.pdf

Pat’s Intel was still foundry first and that is what he came back to do. It wasn’t completely wrong but what it missed was that Intel’s product slate and x86 wasn’t going to return it to leadership. The IFS needed a lot more work to get Nvidia, Qualcomm or Mediatek on board and start getting higher wafer volume. Subsidize customers to be cost competitive versus $100 billion on unneeded fabs.

 

[fastandfurious6]

truth be told Intel Marketing™ is still powerful, general public still believes intel > amd

but too drunk with power it shot its own company lol

most income is not from general public these days

 

[gdansk]

truth be told Intel Marketing™ is still powerful, general public still believes intel > amd

Intel could be shipping broken products for months on end without a fix and people would still recite that Intel is the beacon of stability and problem-free experience.

Like I said earlier there is absolutely no risk of an AMD monopoly forming (unless Intel commits seppuku).

 

[Markfw]

Intel could be shipping broken products for months on end without a fix and people would still recite that Intel is the beacon of stability and problem-free experience.

This has happened in this thread with all the Broken Raptor lakes.

Quoted for truth.

 

[SiliconFly]

.

 

[Meteor Late]

Not if AMD keeps dropping the ball on certain segments, like selling Strix Point for so much money that it only goes into 1500$ laptops.

 

[adroc_thurston]

like selling Strix Point for so much money that it only goes into 1500$ laptops.

a) stop being poor
b) poor people get KRK1

 

[gdansk]

a) stop being poor
b) poor people get KRK1

HX370 1T means it is for "poor" people. x64 is the ghetto. Get out while you still can.

 

[adroc_thurston]

HX370 1T means it is for "poor" people.

Yeah, the rich boys get the -halo with the full LLC.

x64 is the ghetto. Get out while you still can.

this cope again? splendid.

 

[oak8292]

Not if AMD keeps dropping the ball on certain segments, like selling Strix Point for so much money that it only goes into 1500$ laptops.

It is a large ‘low’ volume die. The non-recurring engineering and mask costs needs to be recovered. X86 volumes are not keeping up with ARM volumes and it is a ‘premium’ product with a ‘software moat’. Manufacturing margin and AMD margins keep companies alive.

 

[Meteor Late]

Premium product? software moat? meh
The only premium part will be Strix Halo.

 

[misuspita]

By that logic AMD would have been dead years ago.

AMD didn't have the huge burden of keeping the fabs alive. That's why they survived, they became leaner and could survive. Intel has to change mentality, cut cruft and still have to pay billions upon billions for the fabs even if they just stay idle.

Not if AMD keeps dropping the ball on certain segments, like selling Strix Point for so much money that it only goes into 1500$ laptops.

There are laptops with strix starting at 950euros (so with taxes included, for US members) in EU

 

[gmillers404]

https://twitter.com/x/status/1862509554736800177

Darkmont area efficiency the 24 core compute tile is just 55mm2 with the L2 on 18A😮

Assuming this is correct.

Continued Momentum for Intel 18A

Progress on lead product designs and process readiness is enabling us to bridge from Intel 20A earlier than we’d planned.

www.intel.com

Are these CLF wafer pics?

And assuming that these are pics of clearwater forest. (What else could those be?)

With a simple rotation and crude some photoshop measurements: (Attached picture is different from measurement area)
Die size based from intel photos: 171x125 px
171px*x*125px*x = 55mm2
x = 0.0507 mm/px

Width = 8.669 mm
Height = 6.337 mm
Calculated area: 54.93 mm2 (Close enough)

Single 4C + L2 Darkmont cluster: 49x54px
Width = 2.484 mm
Height = 2.737 mm

Calculated area: ~6.8 mm2

6.8 mm^2

Which is closer to the size of a 4C + L2 Skymont cluster than I expected.
I expected it to have a bigger size to be honest.
Could this mean that 18a has better logic density than N3B?

Darkmont is expected to have more logic components (at least a better and bigger branch predictor from what i've heard) and assuming that the L2 size is the same this could be one of the conclusions i can make.