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.

 

[MS_AT]

It should have stopped at 256-bits. From now the Moore's Law gains are crashing. Going from 100% density gains to 30%. At one point it'll be like everything non-computers, where advancements come hard, expensive, and slow. Like in architecture, energy storage/generation.

Having register at cache line width has its advantages.

I disagree that only showing specific benchmarks where Zen2 wins is the best way to cover Skymont's strengths and weaknesses. Including the 3950X in the main SPEC summary charts would've offered far more context. They could still discuss how poorly Skymont performs in SIMD benchmarks, but they have the added benefit of showing how much Skymont, Zen4, and Zen5 beat Zen2 in the rest of the SPEC suite.

So if I get this right, you would be okay if they compared Skymont SIMD performance to Zen4, since that core was featured in SPEC results in the same article? To be fair I think Zen2 was a good comparison point. It clocks the same, has similar cluster size. It's outdated thought as it comes from 2019, but still thanks to native AVX2, is able to exceed Skymont. That underlines Skymont weak point. They could have used Coffee Lake, but Coffee Lake clocks higher etc. It's just driving home the point the core has a weakness. Since testing SPEC takes a lot of time, I guess they assumed the reader would know that Zen2 rather would be owned by Skymont seeing how much bigger the core is, so they did not show it to save time. But you can always ask the author to be sure. I don't think however it's fair to say they are playing favourites. Just my take on it. You don't have to agree of course

 

[Darkmont]

Hmm...JEDEC must have an upcoming DDR5-7200 spec.

Been in the spec for years.

upgradable DDR5 7200 is rare especially JEDEC

Dell have 16 and 18 inch ARL-HX Pro Max laptops with DDR5-7200 CAMM2. They’ll be in greater volume by the time Panther Lake is fully ramped.

 

[511]

Been in the spec for years.

Dell have 16 and 18 inch ARL-HX Pro Max laptops with DDR5-7200 CAMM2. They’ll be in greater volume by the time Panther Lake is fully ramped.

Panther Lake Fully ramping means Q2 26 best case my guess also when PTL-H fully ramps Intel will minimize selling ARL-H/U

 

[DrMrLordX]

i found it on twitter lol

That first typo . . . "Raper Lake"??? I hope that's a typo. Also what's Amston Lake? Looks like it's involved in edge computing. 2-8c Darkmont or Skymont or . . . ? Twin Lake looks like it's Gracemont so nothing interesting there.

 

[511]

https://twitter.com/x/status/1959895143957561730

SOSA-Aligned VNX+ CPU Blade with Intel® Panther Lake Processor | ADLINK

3U VPX Processor Blades,SOSA-Aligned VNX+ CPU Blade with Intel® Panther Lake Processor

www.adlinktech.com

 

[poke01]

did we see these boards for lunar or arrow lake

 

[511]

did we see these boards for lunar or arrow lake

Nope only MTL afaik ARL -H is not being promoted as Industrial Product

 

[dullard]

View attachment 129194

I'm confused on the tiles here. Do some Panther Lake CPUs now have 6 tiles instead of 5?

vs

 

[511]

I'm confused on the tiles here. Do some Panther Lake CPUs now have 6 tiles instead of 5?
View attachment 129197
vs
View attachment 129198

It says preliminary looks like they forgot to update photos. They are still using MTL Tilesin photos.

 

[511]

 

[ToTTenTranz]

Panther Lake H seems to be pretty decent, especially as AMD has apparently decided to keep the Strix line for what seems to be an eternity.

 

[511]

Crestmont vs Darkmont 17% IPC Increase(seems a bit low)

 

[Philste]

That would mean Darkmont is way slower than Skymont, or is the Crestmont Version of Sierra Forest way faster than desktop crestmont/gracemont?

 

[511]

That would mean Darkmont is way slower than Skymont, or is the Crestmont Version of Sierra Forest way faster than desktop crestmont/gracemont?

I guess it's the later part cause SRF has a generous L3

 

[Saylick]

Here’s some slides:

 

[Magio]

https://twitter.com/x/status/1960012046445011243

So according to this Darkmont IPC would be close to Raptor Cove IPC. If that's correct and it clocks better than Crestmont that would be a decent proposition. Not quite AMD like in pure IPC though.

 

[511]

i was hoping around ~20%

 

[511]

something we missed L2 is 17Cycles vs 19Cycles for Skymont Cluster for 4MB this is a very nice improvement.
This is the same Latency as LNC for L2.

 

[lightisgood]

I got. Darkmont's uarch is roughly equal Skymont's.

I think the improvement of system architecture is good.
A base tile has 192MiB LLC.
This is not so much, however, the load of mesh-interconnect should be crazy reduced.

 

[Josh128]

https://twitter.com/x/status/1960012046445011243

So according to this Darkmont IPC would be close to Raptor Cove IPC. If that's correct and it clocks better than Crestmont that would be a decent proposition. Not quite AMD like in pure IPC though.

Remember when everyone said that Skymont was going to be equal to Raptor Cove IPC? What happened??

 

[AcrosTinus]

Remember when everyone said that Skymont was going to be equal to Raptor Cove IPC? What happened??

It is very close but in SIMD it cannot keep up with a full P-Core.

 

[511]

Remember when everyone said that Skymont was going to be equal to Raptor Cove IPC? What happened??

It is if you read the tweet correctly lol 3.71 and 3.72 also we have to see how Darkmont Performs cause it doesn't have weak L3 Cache like ARL

 

[adroc_thurston]

I got. Darkmont's uarch is roughly equal Skymont's.

I think the improvement of system architecture is good.
A base tile has 192MiB LLC.
This is not so much, however, the load of mesh-interconnect should be crazy reduced.

View attachment 129217

slide says performance and doesn't provide skt rates vs SRF/comp?
lmao Intel.

 

[Covfefe]

https://twitter.com/x/status/1960012046445011243

So according to this Darkmont IPC would be close to Raptor Cove IPC. If that's correct and it clocks better than Crestmont that would be a decent proposition. Not quite AMD like in pure IPC though.

I'd take this tweet with a grain of salt. The 17% number on the Intel slide is SpecIntRate, i.e. one specInt instance for every hardware thread. All the numbers in the tweet (except Darkmont) are specInt.

 

[Josh128]

It is if you read the tweet correctly lol 3.71 and 3.72 also we have to see how Darkmont Performs cause it doesn't have weak L3 Cache like ARL

What do you mean? The tweet clearly says Skymont is 3.18, RPC is 3.71. I remember Intel hyping up "Chadmont" saying it was going to have Raptor Cove +2% IPC. This tweet is saying Darkmont has Skymont+17% IPC and now Darkmont is equal to RPC, lol.