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.

 

[511]

Yeah, but might thread 'eat' the second core cache ? I mean, both core will compte for cache then, no ?
Also, more read/write ports could slow cache access (speed/latency) or augment complexity ?
This might be completely false, I dunno much about cache workings.

It will depend how the cache is shared when mutliple cores access it does each core get a equal amount or not?

 

[DavidC1]

There are rumors that the NEX division will be given up, but the network/wifi I don't think it's possible to let go
I think that the dismantling of the NEX division will be merely a change of personnel within the Intel company.

This looks more than that.

Quote from Reuters:

And Intel’s networking business too may be sold, because companies such as Broadcom have a tight grip on crucial portions of that market, another source said.

Yea, Intel will be a skeleton afterwards.

If they sell the WiFi division, it's going to be a mistake. There is some sort of synergy between products that work together and work well. Their WiFi for example have great performance, have low power, and drivers too. Broadcom may be better than Realtek, but they are not Intel level for client.

First thing I do on an older laptop is upgrade to Intel WiFi, because it actually allows the CPU to reach lower C7/8 power states.

Lip-Bu Tan does not get it. Intel's successes were because they drove the market in standards and compatibility, and that in turn benefitted the PC market.

People probably don't care as much now, but AMD even now doesn't do a fraction of pushing the ecosystem. x86 vs ARM race is ever more closer to reaching the finish line, and it looks like we can discern who the winner is. That is why the DIY market will really die, despite nauseatingly endless claims that "Desktop PC is dead". You'll see when it's really dead.

 

[poke01]

interesting

 

[adroc_thurston]

View attachment 125668

interesting

Not news?

 

[DrMrLordX]

interesting

How old is that? Arrow Lake on 20a clearly didn't happen.

 

[igor_kavinski]

Razer Lake on 14A in 2027 looks a tad bit ambitious, no? Why they skippin' 16A? Found some magic wand to make a bigger node jump from 18A?

 

[poke01]

Not news?

news to me, didn't know razor lake used N2P

 

[Io Magnesso]

Yes, his WiFi product is excellent and he provides a solid driver, so it's easy to use on Linux.
Moreover, Intel is very good at establishing standards. Much better than AMD and ARM.
I can agree with that
However, it is unthinkable to sell network-related products.

 

[511]

View attachment 125668

interesting

He doesn't even know the SKU there is no arrow lake P as well and 20A version got canned 🤣.
Razer Lake S looks like a refresh to me

 

[igor_kavinski]

Razer Lake S looks like a refresh to me

Refresh on better node? So just a straight die shrink?

 

[511]

Refresh on better node? So just a straight die shrink?

i said looks like lol i don't know for sure.
Razer Lake is based on Griffin Cove the last P core before we get Unified Core also Razer Lake also contains LNL Successor RZL-M from this patch

https://twitter.com/x/status/1875246620448846178

 

[MS_AT]

Btw, what is the idea with unified core. Will they go back to using only one core type, or will it be more like Strix Point with two physical implementations of the same core?

 

[511]

looks like it developing multiple core takes lot's of resources. They want to have one core with multiple implementation rather than two separate cores.

 

[Tigerick]

i said looks like lol i don't know for sure.
Razer Lake is based on Griffin Cove the last P core before we get Unified Core also Razer Lake also contains LNL Successor RZL-M from this patch

https://twitter.com/x/status/1875246620448846178

Hmm, instead of PTL-U, now we got PTL-P which should have 28W TDP. Another false claim of LNL efficiency by Intel...

And don't expect similar performance with ARL-H......In fact, Gorgon Point should have slighly faster performance than PTL-H....

 

[511]

Hmm, instead of PTL-U, now we got PTL-P which should have 28W TDP. Another false claim of LNL efficiency by Intel...

TDP != Efficiency also if anything Intel's naming is way better than AMD 🤣 . AMD Naming is cringe.
also on PTL-P isn't that the SKU with 12 Xe3 iGPU?

And don't expect similar performance with ARL-H......In fact, Gorgon Point should have slighly faster performance than PTL-H....

I was not expecting much higher peak MT from PTL tbh 10% above ARL-H in both single and multi will do as long as the battery life is in similar range to LNL.

 

[Io Magnesso]

Hmm, instead of PTL-U, now we got PTL-P which should have 28W TDP. Another false claim of LNL efficiency by Intel...

And don't expect similar performance with ARL-H......In fact, Gorgon Point should have slighly faster performance than PTL-H....

You won't know until you actually see the real thing…
GPUs will be faster

 

[Io Magnesso]

And the Panther Lake-U should still be there

 

[AMDK11]

Is it true that ArrowLake no longer has a ring bus? That it uses mesh entirely?

 

[511]

Is it true that ArrowLake no longer has a ring bus? That it uses mesh entirely?

It still uses ring bus albeit a very slow one at 80 something cycles RING/LLC

 

[Doug S]

Why they skippin' 16A? Found some magic wand to make a bigger node jump from 18A?

Intel never had a 16A, so there's nothing to skip. You're confusing it with TSMC A16.

 

[DavidC1]

You can't trust "AI" for facts, because it often makes stuff up, but with the arrogance that it's completely right, until it's corrected.*

Core logic count couldn't have jumped that much because it's the most power hungry part of the core, and die size shows otherwise. So it took just 33% go Sunny Cove for 20%, but the next iteration took 220% for the same, and the one after that took nearly 100% for just 10%?

Also it calls Skylake "compact and efficient"?

If it was a human, I'd say it's trolling you.

No — there's no clear winner.
Intel’s Lion Cove and AMD’s Zen 5 use two fundamentally different approaches to branch prediction (PBU), each optimized for different execution styles and workloads.

LOL, Lion Cove underperforms in some cases compared to the predecessor. Zen 5 is miles ahead.

 

[poke01]

You can't trust "AI" for facts, because it often makes stuff up, but with the arrogance that it's completely right, until it's corrected.*

Core logic count couldn't have jumped that much because it's the most power hungry part of the core, and die size shows otherwise. So it took just 33% go Sunny Cove for 20%, but the next iteration took 220% for the same, and the one after that took nearly 100% for just 10%?

Also it calls Skylake "compact and efficient"?

If it was a human, I'd say it's trolling you.

LOL, Lion Cove underperforms in some cases compared to the predecessor. Zen 5 is miles ahead.

Everything sucks about Lion Cove, the only saving factor was that it was made on the most dense node. Otherwise, the PPA would have been so embarrassing, a donkey could have done a better job.

 

[Io Magnesso]

Everything sucks about Lion Cove, the only saving factor was that it was made on the most dense node. Otherwise, the PPA would have been so embarrassing, a donkey could have done a better job.

Well, the performance is the worst, but at least I think it should be connected to the bottom...
I hear it's easier to design

 

[Io Magnesso]

Even so, I feel that Lion Cove has become much smaller than Raptor Cove.
I'm not defending Lion Cove...
It's terrible, but compared to Intel's Pentium D and AMD's FX, Compared to these I think the CPU that has not been well received recently is still better.

 

[igor_kavinski]

It's like trying to get a child to think better. What is 2+2? 5? Are you sure? Ummm....6? Think harder!