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.

 

[LightningZ71]

ARL-H is the main volume part of Arrowlake on laptops and it has enough battery endurance to compete with ARL-U plus has lot more OEM designs and volume.


We can pick whatever way but the improvements are to be done and welcome by other user segments. By that line of thinking for many casual users an 8th-10th gen processor laptop with an SSD would suffice.

Considering that most of the machines that we replaced were Whiskey Lake and Tiger Lake U/P going to Meteor Lake, aside from the system no longer getting bogged down for a few minutes twice a day with audit cycles, no one really noticed much of an improvement while on their docks.

This has more to do with an earlier assertion of mine that, aside from us throwing more background stuff on computers for security, computers years ago reached a level of "enough performance" for the vast majority of "office" users.

 

[511]

https://twitter.com/x/status/1972360365435650401

 

[Doug S]

Bruh using MLID as a source means the Intel engineer was a vaporware.

Well more to the point, how is MLID going to know some guy emailing or DMing him is an Intel engineer? He's not going to contact him via his @intel.com email address or send him a scan of his badge. Heck even if he did it is not that difficult to fake either if you wanted to wind MLID up.

That's a hazard with anyone receiving leaked information I suppose, but if you had met a guy at a conference or something where he was representing that company and you exchanged contact info you would know he's who he presents himself as. Or if several times he'd leaked new information that later was proven 100% accurate, you could infer he's either who he claims to be or is acting as a go-between for the legit person.

 

[Josh128]

He says "lower expectations", but I havent really seen any hype about PTL regarding clocks.

https://twitter.com/x/status/1972149206354911619

 

[poke01]

He says "lower expectations", but I havent really seen any hype about PTL regarding clocks.

https://twitter.com/x/status/1972149206354911619

The most exciting thing about Panther lake is made by TSMC. The CPU is boring.

 

[Josh128]

The most exciting thing about Panther lake is made by TSMC. The CPU is boring.

Disagree. Even if the CPU is complete shat, its our first glimpse at the legendary 18A on which the company was bet.

 

[poke01]

Disagree. Even if the CPU is complete shat, its our first glimpse at the legendary 18A on which the company was bet on.

The node is also bad. If that’s what Pat bet the company on, you can see how bad the management was.

 

[Khato]

And that's how Intel lost BILLIONS!

Doesn't matter what he was for or against. What matters in the end is actions he took and decisions he made. And almost everyone agrees that he's the most unexpected brilliant idiot CEO Intel has had. He was supposed to be the chosen one.

We have an Intel engineer telling MLID that cancelling Royal Core was probably a big mistake. Who made that decision?

Heh, I'll believe actual Intel engineers rather than MLID sources. I'd compare the work on Royal Cove to more of an academic research project than an actual product design cycle. What exactly was Pat supposed to do after years of development with nothing to show for it? Cancelling the paid holiday and dismantling the teams appropriately was the only rational decision to be made.

My understanding is that the only shares repurchased under Pat's tenure were in Q1 of 2021 under terms that were in place before he took on the role. So no, unlike previous CEOs Pat did not waste billions of dollars on repurchasing stock. Instead he invested in employees and assets. It's fair to argue that he invested a bit too aggressively in each...

I'd argue that his primary failing was not taking the axe to Royal Cove and all the other middle to upper management that were feeding him unrealistic projections and schedules sooner. It remains to be seen if LBT is actually fixing that issue or just reducing the number of layers without addressing the root problem.

And if the current CEO succeeds, people will say, oh, it was all Pat's strategy. LBT simply executed it blindly without changing a thing!

Yeah right. Probably why Pat was fired because the only thing he couldn't do was execute. You know, something good CEOs are supposed to do?

If LBT succeeds it'll only be because he was finally forced into playing politics. Both political parties in the US recognize the need to incentivize semiconductor manufacturing in the US. That's by far the best factor in Intel's favor at the moment and the fact that LBT didn't start working it on day one is just one of many indicators of his incompetence.

And I can't help but laugh at the notion that CEOs are good at execution. They're nothing more than the cherry on top of the management structure. They take data in from their reports, make decisions based on that data, and then tell those reports what to do.

 

[poke01]

And I can't help but laugh at the notion that CEOs are good at execution. They're nothing more than the cherry on top of the management structure

Depends, most founder CEOs are good and make their own decisions

 

[CouncilorIrissa]

He says "lower expectations", but I havent really seen any hype about PTL regarding clocks.

https://twitter.com/x/status/1972149206354911619

He's referring to a certain twitter schizo spreading 5.5-5.7 GHz figures. Not linking the post so as not to give him any exposure.

The node is also bad. If that’s what Pat bet the company on, you can see how bad the management was.

How about we wait for the chip to come out before proclaiming how bad it is.

 

[poke01]

How about we wait for the chip to come out before proclaiming how bad it is

Okay, I’ll shut up till then

 

[Doug S]

Depends, most founder CEOs are good and make their own decisions

I think a lot of people would take issue with the claim that "most founder CEOs are good". Yes they make their own decisions but that's usually because they own so much stock they are effectively untouchable, and people buy (or decide not to buy) the stock knowing you're getting that founder for better or for worse.

Yes there are stellar examples like Jensen Huang, and there are examples of people who learned the hard way like Steve Jobs, but many founder CEOs are not at all good. You can't always tell because how do you know whether a company is successful because of its leadership or despite its leadership. Is Zuckerberg a good CEO? Meta has been very successful, but even the name "Meta" brings to mind some of his stupenously bad decisions that cost the company many tens of billions. We don't have an alternate universe to test theories like "what if Facebook had replaced Zuckerberg 15 years ago" or "where would Apple be today if Steve Jobs hadn't returned with his lessons from his Apple firing and founding of NeXT and Pixar under his belt".

Developing great technology takes one skillset - being great with software or hardware. Think Steve Wozniak. Founding a company takes another skillset - you have to be a bit of a risk taker, and have some real people skills to sell investors on your dream unless through luck of birth or circumstance you are rich enough to finance it all yourself. Steve Jobs to a T. Running a company takes a third skillset - people management is something most engineers suck at, you have to find a way to connect your products to people who can benefit from them, and you have a lot of financial knowledge to properly manage your company's growth.

There are many people with one of those skillsets - including many reading this. There are a smaller number with two but they aren't terribly rare. People with all three are unicorns.

 

[Kepler_L2]

How about we wait for the chip to come out before proclaiming how bad it is.

We know enough to confirm 18A sucks. It's worse than N3P despite GAAFET + BSPDN

 

[511]

We know enough to confirm 18A sucks. It's worse than N3P despite GAAFET + BSPDN

Not really it's Fmax is not that good a '+' would fix it but the curse of 5.1 GHz is real though
TGL/MTL/PTL all at 5.1 Ghz

 

[poke01]

@igor_kavinski , oct 9 is the date

 

[511]

So I assume the NDA Date was not under NDA?

 

[igor_kavinski]

It kinda sucks that I waited two weeks and now 10 days more waiting.

 

[511]

https://twitter.com/x/status/1972630014584832036

 

[poke01]

screenshot

https://twitter.com/x/status/1972630014584832036

 

[branch_suggestion]

https://twitter.com/x/status/1972630014584832036

It better be in good volume at CES since TTM is the only saving grace of this thing.
That and the internal node usage but 18A isn't even cheap for Intel.

 

[Josh128]

We know enough to confirm 18A sucks. It's worse than N3P despite GAAFET + BSPDN

$INTC to the moon!!

 

[poke01]

Is the GPU in PTL using FinFlex?

 

[Win2012R2]

Not really it's Fmax is not that good

Which might just be enough, and who says there are no other problems with yields/cost too?

Well more to the point, how is MLID going to know some guy emailing or DMing him is an Intel engineer?

He asks for a photo made in Intel's toilet whilst holding current newspaper...

So I assume the NDA Date was not under NDA?

The mere existence of NDA is under NDA...

 

[Josh128]

View attachment 131196

screenshot

No way man. Is this thing going to be Cannon Lake / Ice Lake 2.0?? ie an "investor requirement" launch??

 

[511]

Which might just be enough, and who says there are no other problems with yields/cost too?

It's literally their best yielding product in the last 10 years though the bar is not high

No way man. Is this thing going to be Cannon Lake / Ice Lake 2.0?? ie an "investor requirement" launch??

How is it Cannon Lake/Ice Lake 2.0? Those didn't even go over 4 GHz even on a 4 wide core this is 5 GHz 8 wide LNC+ it's EEP Intel has done this for like quite some time.

He asks for a photo made in Intel's toilet whilst holding current newspaper...

Maybe he works as a janitor part time for the companies 🤣🤣.