• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads

Page 990 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
T

Tigerick

Senior member
Wildcat Lake (WCL) Specs

Intel Wildcat Lake (WCL) is upcoming mobile SoC replacing Raptor Lake-U. 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 Q1 2026.

Intel Raptor Lake UIntel Wildcat Lake 15WIntel Lunar LakeIntel Panther Lake 4+0+4
Launch DateQ1-2024Q2-2026Q3-2024Q1-2026
ModelIntel 150UIntel Core 7 360Core Ultra 7 268VCore Ultra 7 365
Dies2223
NodeIntel 7 + ?Intel 18-A + TSMC N6TSMC N3B + N6Intel 18-A + Intel 3 + TSMC N6
CPU2 P-core + 8 E-cores2 P-core + 4 LP E-cores4 P-core + 4 LP E-cores4 P-core + 4 LP E-cores
Threads12688
Max Clock5.4 GHz4.8 GHz5 GHz4.8 GHz
L3 Cache12 MB6 MB12 MB12 MB
TDP15 - 55 W15 - 35 W17 - 37 W25 - 55 W
Memory128-bit LPDDR5-520064-bit LPDDR5x-7467128-bit LPDDR5x-8533128-bit LPDDR5x-7467
Size96 GB48 GB32 GB128 GB
Bandwidth83 GB/s60 GB/s136 GB/s120 GB/s
GPUIntel GraphicsIntel GraphicsArc 140VIntel Graphics
RTNoNoYESYES
EU / Xe96 EU2 Xe8 Xe4 Xe
Max Clock1.3 GHz2.6 GHz2 GHz2.5 GHz
NPUGNA 3.017 TOPS48 TOPS49 TOPS






PPT1.jpg
PPT2.jpg
PPT3.jpg



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.



LNL-MX.png
 

Attachments

  • PantherLake.png
    PantherLake.png
    283.5 KB · Views: 24,049
  • LNL.png
    LNL.png
    881.8 KB · Views: 25,534
  • INTEL-CORE-100-ULTRA-METEOR-LAKE-OFFCIAL-SLIDE-2.jpg
    INTEL-CORE-100-ULTRA-METEOR-LAKE-OFFCIAL-SLIDE-2.jpg
    181.4 KB · Views: 72,443
  • Clockspeed.png
    Clockspeed.png
    611.8 KB · Views: 72,329
Last edited:
Covfefe said:
I'm not so sure about that. AFAIK Wildcat doesn't have a low power island / lpe cores. Also Alder Lake-N used chiplets without advanced packaging and Wildcat Lake might follow suit to keep the price down.
Click to expand...
Yea like @511 says Wildcatlake's E cores are really LPE cores, it has a 4MB memory side cache, and UCI-E. It's a low cost Pantherlake.
igor_kavinski said:
Atom, N-series, Pentium...Intel destroyed their brand name with those hideous products that must have contributed to hundreds of thousands of tons of toxic silicon waste in landfills. An image conscious company like Apple would never put out products like that.
Click to expand...
Actually the N-series are great for what they are. There's one single board computer with N100 that uses 1.5W while idle, and has NVMe for storage and SO-DIMM for RAM, meaning upgradeability. There's also a miniPC for $129. You have tons of other options with eDP, multiple ethernet and USB, for a home router. They are essentially Skylake class performance cores so performance is not too shabby.

Go put Linux and there's no platform that has the perfect balance of being open, and with wide compatibility with everything.
 
Last edited:
511 said:
the issue is IFS getting customer otherwise Intel Foundry is over
Click to expand...
I have the solution to that.

Create a simple human sized robot with multiple compartments for holding stuff and groceries etc. that follows the owner around like a faithful dog. The SoC for that is going to be enough to fill the fabs. Pretty soon, world population will double, with half of them being grocery carrying robots!
 
511 said:
the issue is IFS getting customer otherwise Intel Foundry is over
Click to expand...

Wouldn't Intel Design be over too? They have a node advantage over AMD and yet have about ~35% worse power efficiency with their TSMC designed chips. Even if all they did was become a design firm they would keep losing margins to AMD over time due to inefficient designs. If anything I would say Intel Foundry is miles ahead of Intel Design at this point.
 
therealsonnen said:
Wouldn't Intel Design be over too? They have a node advantage over AMD and yet have about ~35% worse power efficiency with their TSMC designed chips. Even if all they did was become a design firm they would keep losing margins to AMD over time due to inefficient designs. If anything I would say Intel Foundry is miles ahead of Intel Design at this point.
Click to expand...
The biggest problem with this conclusion is that Inttel design makes money. IFS does not. The hope is that IFS will soon break even later this year (if 14A is canned and 18A doesn't get any external customers), with Intel design carrying their butts to do so.
The design side is way less costly to fix and run than the foundry side as well.
If Intel foundry is over, the design side will prob lose a good bit of market share as they slowly shift over to tsmc, but by no means is it a death sentence.
 
Geddagod said:
The biggest problem with this conclusion is that Inttel design makes money. IFS does not. The hope is that IFS will soon break even later this year (if 14A is canned and 18A doesn't get any external customers), with Intel design carrying their butts to do so.
Click to expand...

At one point, the most optimistic outlook was for the foundry to break even in 2027, which may have been delayed further out.

It is going to get worse (through first 2026 quarters) before it gets better, because ramping new process typically has higher costs associated with it.

Geddagod said:
The design side is way less costly to fix and run than the foundry side as well.
If Intel foundry is over, the design side will prob lose a good bit of market share as they slowly shift over to tsmc, but by no means is it a death sentence.
Click to expand...

Probably the biggest difference is that if the design side has a good design, it can be turned over to TSMC, and you have large quantity of the actual chips being churned out by TSMC in no time.

When the foundry side gets a good "recipe", it takes 10s of billions in investments, and years of construction before foundry can make a profit.

Because of this, if there is to be a success with Intel foundry, it will be years down the line.
 
therealsonnen said:
Wouldn't Intel Design be over too? They have a node advantage over AMD and yet have about ~35% worse power efficiency with their TSMC designed chips. Even if all they did was become a design firm they would keep losing margins to AMD over time due to inefficient designs. If anything I would say Intel Foundry is miles ahead of Intel Design at this point.
Click to expand...
IFS Provides the volume everyone is constrained at TSMC except Intel they have their own fabs and they are ramping that for themselves why do you think AMD hasn't been able to gain share as quickly vs Intel in client/Server.
Geddagod said:
My bad, thought the original claim was that it would break even this year.
Click to expand...
Joe NYC said:
At one point, the most optimistic outlook was for the foundry to break even in 2027, which may have been delayed further out.

It is going to get worse (through first 2026 quarters) before it gets better, because ramping new process typically has higher costs associated with it.
Click to expand...
I think Dave said they haven't changed their guide for 2027 break even but they have said that if external customer arrives it might get delayed due to additional spending.
 
gdansk said:
I think that belongs in this thread instead.
But other than the 1 Xe cut (can that even do desktop graphics? lol) that looks pretty good
Click to expand...
Even 1XE should be faster than current intel HD (older iris) so it's significant improvement for intended market.
 
hemedans said:
Even 1XE should be faster than current intel HD (older iris) so it's significant improvement for intended market.
Click to expand...
Current one is in Arrowlake with 4 Alchemist+ generation Xe cores. Xe3 is way better but with only one of them I'm not sure how much of an upgrade that will be. And in laptops Arrowlake-H has 8 Xe cores.

Against low end Raptorlake it has 64EUs. It really needs the 2 Xe core version to be a noticeable upgrade.

Wildcatlake is an overall win though. A great part. Also brings 10+ hour battery life easily to budget notebooks. Would love to see pleothera of SBCs based on them.
 
Last edited:
  • Like
Reactions: 511
DavidC1 said:
Current one is in Arrowlake with 4 Alchemist+ generation Xe cores. Xe3 is way better but with only one of them I'm not sure how much of an upgrade that will be. And in laptops Arrowlake-H has 8 Xe cores.

Against low end Raptorlake it has 64EUs. It really needs the 2 Xe core version to be a noticeable upgrade.

Wildcatlake is an overall win though. A great part. Also brings 10+ hour battery life easily to budget notebooks. Would love to see pleothera of SBCs based on them.
Click to expand...
I don't think it replace arrowlake, it's direct replacement to something like intel core 3 100U, i3 1215U, 1315U etc.
 
I disagree about WCL being good, yeah it is cheap, but I hated how Mendocino onlly had 2 CU. It's just not acceptable in 2026 IMO, yeah I understand 2 Xe3 cores will be better than 2 CU from your RDNA2 or 3.5, but still, 4 Xe cores were the way to go, one third the size of the top of the line iGPU is enough, no need for more. Maybe the lowest versions like the Core 3, ok 2 Xe as a super cheap option, but the others, just no.
 
Meteor Late said:
I disagree about WCL being good, yeah it is cheap, but I hated how Mendocino onlly had 2 CU. It's just not acceptable in 2026 IMO, yeah I understand 2 Xe3 cores will be better than 2 CU from your RDNA2 or 3.5, but still, 4 Xe cores were the way to go, one third the size of the top of the line iGPU is enough, no need for more.
Click to expand...
It's targeting N100 replacement and 1215U replacement they also have 2Xe3 Cores and adding extra increases cost
 
511 said:
it only got a single tile of NPU sadly
Click to expand...
A single tile is one too many, either minimize area to save costs or bring performance where it matters with 4 Xe like @Meteor Late suggested.

I can respect either choice but not this "a mix of everything" which looks like it tries to check boxes instead of nailing a clear role for the chips.
 
coercitiv said:
A single tile is one too many, either minimize area to save costs or bring performance where it matters with 4 Xe like @Meteor Late suggested.

I can respect either choice but not this "a mix of everything" which looks like it tries to check boxes instead of nailing a clear role for the chips.
Click to expand...
Maybe with WCL refresh but i don't make the decision i would have removed the NPU all together if i could and made it cheaper. Intel even cut the RT Cores in WCL to save area
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top