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]

In desktop and mobile laptop space, I think Intel Lunar Lake and Arrow Lake will beat Zen 5. I think Arrow Lake slaughters vanilla Zen 4 and 5 in gaming and trades blows with 7800X3D/9800X3D in lightly threaded games and pulls ahead of the 8 core X3D counterparts in heavily threaded games. And I think Arrow Lake beats 12 and 16 core Zen 4 and 5 X3D parts in heavily threaded games mostly.

I can't say about gaming due to ths tiles will add additional latency to IMC <-> CPU but iy should be on par with 7800X3D will definitely loose to 9800X3D

In the server and enterprise space, I think Zen 5 overall wins over Intel Xeon counterparts.

At least Intel will not bleed money like they used to on server for past few years these are actually good in god knows when wendell called it nehlam amd said they should have done it years ago at least Pat's Plan is working somewhat better than last 3 CEOs gotta give him credit

 

[cannedlake240]

Intel will not bleed money like they used to on server for past few years

If Turin is 30% faster(clocks+IPC) vs intel instead of 50% faster as in the past, will that really change things for Intel? GNR does have a few advantages but AMD also has Vcache that's very useful for many workloads. Also GNR cost vs EMR is important, GNR uses a lot of euv silicon but emr's 2 750mm2 on Intel 7 tiles are also quite expensive...

 

[9949asd]

I can't say about gaming due to ths tiles will add additional latency to IMC <-> CPU but iy should be on par with 7800X3D will definitely loose to 9800X3D

At least Intel will not bleed money like they used to on server for past few years these are actually good in god knows when wendell called it nehlam amd said they should have done it years ago at least Pat's Plan is working somewhat better than last 3 CEOs gotta give him credit

What if 9800x3d only 3% faster than 78x3d. 97x is 3% faster than 77x.

 

[511]

If Turin is 30% faster(clocks+IPC) vs intel instead of 50% faster as in the past, will that really change things for Intel? GNR does have a few advantages but AMD also has Vcache that's very useful for many workloads. Also GNR cost vs EMR is important, GNR uses a lot of euv silicon but emr's 2 750mm2 on Intel 7 tiles are also quite expensive...

It won't be 30% tbf on average it will be lot close than that there is also "AI" which AMD will undoubtedly loose to no matter how much they say AMD will trounce intel in NAMD for sure though we will see on 11 October though

 

[dullard]

I wonder why Intel won't do 3D vcache they have the tech now with Intel 3T and they are using it in Clearwater Forest

I've been wondering that for years. Intel announced it at CES 2019 (with hints about it throughout 2018), years before AMD had 3D cache. But, Intel never released it.

 

[511]

I've been wondering that for years. Intel announced it at CES 2019 (with hints about it throughout 2018), years before AMD had 3D cache. But, Intel never released it. View attachment 108833

Yes at least bring back broadwell L4 cache

 

[MS_AT]

Yes at least bring back broadwell L4 cache

I think Intel is going the other way around, since they have introduced L0 cache with LionCove, I guess any stacked cache will be L-1, especially if they would put it at the bottom, then it would be in a basement, so to speak. Though not sure if basement cache would be a good marketing name...

 

[511]

Grandma's Basement Cache worth $700k

 

[AMDK11]

https://twitter.com/x/status/1843334836322537916

 

[Hitman928]

https://twitter.com/x/status/1843334836322537916

The OCD part of my brain hates it, lol.

Edit: I am surprised there isn't a dummy tile added at the top for structural support.

 

[Hulk]

I'm showing static clocks @ 5400mhz (at 250w)

What you see in reviews are autoboosting at 200w (stock PPT limit for the 9950X) --> real clockspeed @ stock is only ~4800mhz in Cinebench R23
And like i have been saying from the start, vanilla Z5 stock V/F curve is pretty bad as it leaves alot on the table

Even leaks are showing the 9950X3D getting higher MT score with same PPT limit as 9950X ((even to a larger degreee then what we saw with Zen4)and now the X3D models scale all the way up the power curve)

Wow. I had no idea Zen 5 was clocked so conservatively out of the box.

 

[coercitiv]

I guess any stacked cache will be L-1, especially if they would put it at the bottom, then it would be in a basement, so to speak. Though not sure if basement cache would be a good marketing name...

Sub-cache, the obvious acronym for subway cache!

 

[Thibsie]

Sub-cache, the obvious acronym for subway cache!

Subway cache. That should stick IMO

 

[Hulk]

Fun with Excel.

 

[Hulk]

I think some of my Zen data might not be accurate. If anybody has Zen<5 and want to test please post scores and average clock during the run. I can't let go of these pet projects until they are finished.

 

[Abwx]

Zen 4 does 482 pts/GHz/core with SMT, in ST that s 371 pts/GHz

The 7700X run at 5.3 in MT and does 20437 pts in Computerbase review, they measured almost 5.3 in Blender, your scores would imply that it run at roughly 5.65 all cores and that the SMT gain is only 23%, and also that Zen 5 best it by 21.9% better IPC in MT.

About all those CPUs ST and MT scores, excluding Zen5, can be found here, the frequencies measurements are scattered in the respective CPUs reviews.

Intel Core i9-14900KS im Test: Anwendungs-Benchmarks, Leistungsaufnahme & Temperatur

Intel Core i9-14900KS im Test: Anwendungs-Benchmarks, Leistungsaufnahme & Temperatur / Testsystem und Testmethodik für Anwendungen

www.computerbase.de

 

[Hulk]

 

[511]

Grandma's Basement Cache worth $700k

No one got the joke except one person 🥲

 

[jdubs03]

Saw this just now from HXL:

https://twitter.com/x/status/1843538420846735394

Hopefully, we’ll get an English version soon.

The commentary to that post are all saying like 2% decrease in gaming performance.

 

[511]

Chadmont 32% 🤣

 

[Det0x]

wow if true

6200/5700 -> ~9% clockspeed regression
+9% IPC for "Gen15" vs Gen14

End results is that ST performance dont really change that much (?)
Will also be interesting to see how this uncore perform in gaming, so far we have seen not too good latency for Arrow Lake

 

[coercitiv]

Will also be interesting to see how this uncore perform in gaming, so far we have seen not too good latency for Arrow Lake

The uncore is probably the culprit for the lower IPC uplift (desktop relative to mobile).

 

[jdubs03]

Wait there’s more… courtesy of wxnod on Twitter.
Showing a 3% lower result in GB6.3 ST vs. the 9950X.
Looking pretty meh.

 

[coercitiv]

 

[poke01]

https://twitter.com/x/status/1843547171511251341

Intersting how the 9950X is ahead in Geekbench