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.

 

[Khato]

A bug whose correction improve perf/watt by 20%..?..
And you think that Intel engineers were unaware of the thing till launch day..?.

Yup.

Two distinct possibilities as to why the early review samples had months old pcode. First is that's the pcode version manufacturers were originally supplied with when starting their product validation and no one told them that they needed to update before sending out the review samples. Second is that it was intentional - get one mixed review cycle on launch, then get a second thanks to this large efficiency improvement as more laptop designs become available.

 

[Hulk]

Slay 8 GB RAM!

32GB is the new 8GB.

 

[Geddagod]

Here we have some die size estimations.

https://twitter.com/x/status/1736059767880085868

Twitter link doesn't work for some reason, it's from Andreas Schilling.

pretty much the same area as a Zen 4 CCD. cores are larger, but the L3 cache and ringbus elements, as well as the interconnect overhead are prob smaller relatively.

 

[Hulk]

Meteor Lake in the Ark

No base frequency spec for base power, which is strange?
Max frequency for SoC island CPU's is 2.5GHz.
185H being listed as launched Q4 '24. Really?

Products formerly Meteor Lake

ark.intel.com

 

[TESKATLIPOKA]

32GB is the new 8GB.

Considering a lot of laptops have non-removable memory, It's stupid to buy anything less than 16GB, honestly, even that is not much If you have an IGP and want the laptop to be at least a bit future-proof.

Meteor Lake in the Ark

No base frequency spec for base power, which is strange?
Max frequency for SoC island CPU's is 2.5GHz.
185H being listed as launched Q4 '24. Really?

Products formerly Meteor Lake

ark.intel.com

I see there Q4'23.

 

[Saylick]

32GB is the new 8GB.

Jensen, you hearing these words??

 

[DrMrLordX]

It would seem that some people are still clueless about how PL1/2 and Tau works on Intel mobile CPUs since Tiger Lake. As a result, they clutter these Intel threads with useless information and analyses.

It would seem that if Intel releases a product that does not hold up to the same testing regime that has been used for mobile CPUs for years, that Intel has a problem. Not the reviewers.

 

[TESKATLIPOKA]

Jensen, you hearing these words??

Jensen is too busy counting his money.

 

[Glo.]

32GB is the new 8GB.

ESPECIALLY for unified memory architectures.

 

[Saylick]

Jensen is too busy counting his money.

For him, 32B is the new 8B (USD).

 

[Khato]

For all the noise that both Intel and AMD wish to make about the AI PC, they really need to hand-hold some reviewers or something to get such included in benchmark suites. The notebookcheck.com variant of their review included the UL Procyon AI Inference benchmark run on both GPU and NPU to compare performance and power, but there aren't any other comparison points of interest. (Summary of those results by the way is GPU is around 10% faster than NPU, but basically twice the power.) Have there been any reviews of MTL or Phoenix that properly dig into the NPU performance and efficiency? Or are we just supposed to take Intel's and AMD's word that these are magical IP blocks because they can be used for AI?

 

[controlflow]

Meteor Lake in the Ark

No base frequency spec for base power, which is strange?
Max frequency for SoC island CPU's is 2.5GHz.
185H being listed as launched Q4 '24. Really?

Products formerly Meteor Lake

ark.intel.com

Golden Pig already has the 185H in his benchmarks. No way Q4 24 is right.

 

[controlflow]

For all the noise that both Intel and AMD wish to make about the AI PC, they really need to hand-hold some reviewers or something to get such included in benchmark suites. The notebookcheck.com variant of their review included the UL Procyon AI Inference benchmark run on both GPU and NPU to compare performance and power, but there aren't any other comparison points of interest. (Summary of those results by the way is GPU is around 10% faster than NPU, but basically twice the power.) Have there been any reviews of MTL or Phoenix that properly dig into the NPU performance and efficiency? Or are we just supposed to take Intel's and AMD's word that these are magical IP blocks because they can be used for AI?

Have not seen much beyond this.

The claim is that CPU is best for low latency. GPU wins for throughput and NPU wins for efficiency/lowest energy cost for completing the task.

 

[IEC]

https://www.bestbuy.com/site/asus-zenbook-14-oled-14-3k-touch-laptop-intel-core-ultra-7-32gb-memory-1tb-ssd-ponder-blue/6569307.p?skuId=6569307

This one is 32GB of LPDDR5 7466 according to this review https://www.ultrabookreview.com/66513-asus-zenbook-14-ux3405-review/

Careful there, the BestBuy full specs list says:
"System Memory (RAM)

32 gigabytes
Type of Memory (RAM)

LPDDR5X
System Memory RAM Speed

6400 megahertz"

Wouldn't be the first time a brand shipped its fastest configuration for the review only to quietly downgrade components for the mass market version hoping no one notices.

 

[controlflow]

Careful there, the BestBuy full specs list says:
"System Memory (RAM)

32 gigabytes
Type of Memory (RAM)

LPDDR5X
System Memory RAM Speed

6400 megahertz"

Wouldn't be the first time a brand shipped its fastest configuration for the review only to quietly downgrade components for the mass market version hoping no one notices.

The Best Buy spec list is probably wrong. Up until today it also incorrectly listed the display as an IPS panel (corrected to OLED now) and it still claims it has Intel UHD graphics. It also had the wrong processor name yesterday and they updated that too.

Here is the official specs from Asus

ASUS Announces All-New Zenbook 14 OLED (UX3405)

press.asus.com

The memory claims "Up to 32 GB LPDDR5x 7467 MHz RAM". I suppose the "up to" implies there can be some slower configurations...

 

[tamz_msc]

It would seem that if Intel releases a product that does not hold up to the same testing regime that has been used for mobile CPUs for years, that Intel has a problem. Not the reviewers.

"Testing regime that has been used for mobile CPUs for years" - testing regime that is basically Cinebench on loop?

 

[FlameTail]

ESPECIALLY for unified memory architectures.

AAPL, you heard that?

 

[Hulk]

Golden Pig already has the 185H in his benchmarks. No way Q4 24 is right.

My mistake.

 

[Hulk]

MLID prognostications from May.

Just about everything was wrong.

Computer Architecture

Launches Q3 as 14th gen on desktop and 14th gen "HX" on laptop
5.4GHz boost for 45W models and 5.2GHz boost for 28W "H" Models
12-21% IPC increase over Raptor Lake
6+6 and 2+8 dies
Project IPC increase for Crestmont is +8% over Gracemont
Up to 10% IPC increase over Raptor Lake
512MB Adamantine prototypes being tested
30 to 45% less power at ISO frequency compared to Raptor Refresh below 45W
ARL and MTL share the same base tile so Intel could launch ARL with last gen Battlemage graphics if Lion Cove is ahead of schedule.

 

[controlflow]

MLID prognostications from May.

Just about everything was wrong.

Computer Architecture

Launches Q3 as 14th gen on desktop and 14th gen "HX" on laptop
5.4GHz boost for 45W models and 5.2GHz boost for 28W "H" Models
12-21% IPC increase over Raptor Lake
6+6 and 2+8 dies
Project IPC increase for Crestmont is +8% over Gracemont
Up to 10% IPC increase over Raptor Lake
512MB Adamantine prototypes being tested
30 to 45% less power at ISO frequency compared to Raptor Refresh below 45W
ARL and MTL share the same base tile so Intel could launch ARL with last gen Battlemage graphics if Lion Cove is ahead of schedule.

It really amazes me that the nonsense this clown spews gets so much coverage and stated as news or facts by so many outlets. He is a random guy making YouTube videos out of his attic and his videos make it abundantly clear that he doesn't even have a technical understanding of any of the topics he tries to cover.

 

[Saylick]

It really amazes me that the nonsense this clown spews gets so much coverage and stated as news or facts by so many outlets. He is a random guy making YouTube videos out of his attic and his videos make it abundantly clear that he doesn't even have a technical understanding of any of the topics he tries to cover.

True, but what is also true is that the current state of tech journalism, especially as it pertains to semiconductors, is pretty bad. It's comprised of a bunch of people who don't have a formal background in electrical engineering or software engineering regurgitating whatever is currently trending on the internet, namely Twitter leaks, or just acting as a corporate mouthpiece by making puff pieces. Who cares about accuracy and legitimacy when pumping out crummy articles to rake in ad revenue trumps all? If a crappy site like WTFTech can succeed in today's tech journalism market by purely churning out articles faster and in higher quantities than its competitors, that says a lot imo. The authors there literally were just a bunch of college hacks who turned a hobby into a full-time hustle. I'll give them credit for working harder than the rest, but the bar is super low. I'd argue there's easily dozens of enthusiasts on AT Forum who are more knowledgeable than MLID or the authors at click-bait tech sites, but the reason why we don't make our own website is that we simply have better things to do (like typing out a long rant about said websites ).

 

[FlameTail]

MLID prognostications from May.

Just about everything was wrong.

Computer Architecture

Launches Q3 as 14th gen on desktop and 14th gen "HX" on laptop
5.4GHz boost for 45W models and 5.2GHz boost for 28W "H" Models
12-21% IPC increase over Raptor Lake
6+6 and 2+8 dies
Project IPC increase for Crestmont is +8% over Gracemont
Up to 10% IPC increase over Raptor Lake
512MB Adamantine prototypes being tested
30 to 45% less power at ISO frequency compared to Raptor Refresh below 45W
ARL and MTL share the same base tile so Intel could launch ARL with last gen Battlemage graphics if Lion Cove is ahead of schedule.

You forgot to mention that he retracted a lot of this stuff later, in the months leading up to MTL launch.

 

[Hulk]

You forgot to mention that he retracted a lot of this stuff later, in the months leading up to MTL launch.

Yeah so when you're far out from launch just make things up. As actual information becomes available change your earlier predictions, which were of course how you got the views in the first place. Brilliant and perfectly circular.

 

[SiliconFly]

It really amazes me that the nonsense this clown spews gets so much coverage and stated as news or facts by so many outlets. He is a random guy making YouTube videos out of his attic and his videos make it abundantly clear that he doesn't even have a technical understanding of any of the topics he tries to cover.

The last one is very true. He seriously lacks technical understanding.

 

[FlameTail]

Yeah so when you're

Hol up. I am not MLID !