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

Tigerick · Aug 22, 2022

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.

Comparison of upcoming Intel's U-series CPU: Core Ultra 100U, Lunar Lake and Panther Lake

Model	Code-Name	Date	TDP	Node	Tiles	Main Tile	CPU	LP E-Core	LLC	GPU	Xe-cores
Core Ultra 100U	Meteor Lake	Q4 2023	15 - 57 W	Intel 4 + N5 + N6	4	tCPU	2P + 8E	2	12 MB	Intel Graphics	4
?	Lunar Lake	Q4 2024	17 - 30 W	N3B + N6	2	CPU + GPU & IMC	4P + 4E	0	8 MB	Arc	8
?	Panther Lake	Q1 2026 ?	?	Intel 18A + N3E	3	CPU + MC	4P + 8E	4	?	Arc	12

Comparison of die size of Each Tile of Meteor Lake, Arrow Lake, Lunar Lake and Panther Lake

	Meteor Lake	Arrow Lake (20A)	Arrow Lake (N3B)	Arrow Lake Refresh (N3B)	Lunar Lake	Panther Lake
Platform	Mobile H/U Only	Desktop Only	Desktop & Mobile H&HX	Desktop Only	Mobile U Only	Mobile H
Process Node	Intel 4	Intel 20A	TSMC N3B	TSMC N3B	TSMC N3B	Intel 18A
Date	Q4 2023	Q1 2025 ?	Desktop-Q4-2024 H&HX-Q1-2025	Q4 2025 ?	Q4 2024	Q1 2026 ?
Full Die	6P + 8P	6P + 8E ?	8P + 16E	8P + 32E	4P + 4E	4P + 8E
LLC	24 MB	24 MB ?	36 MB ?	?	8 MB	?
tCPU	66.48
tGPU	44.45
SoC	96.77
IOE	44.45
Total	252.15

Intel Core Ultra 100 - Meteor Lake

As mentioned by Tomshardware, TSMC will manufacture the I/O, SoC, and GPU tiles. That means Intel will manufacture only the CPU and Foveros tiles. (Notably, Intel calls the I/O tile an 'I/O Expander,' hence the IOE moniker.)

Ghostsonplanets · Apr 8, 2024

Intel was claiming 40% scenario power reduction over MTL, alongside others power gating and limited power usage improvements. So parity with best in class Arm designs is certainly achievable.

uzzi38 · Apr 8, 2024

SpudLobby said:
I’ll let Uzzi for ex come in on this if he wants to.

I mean that's what the other forum dweller here has said but I can't confirm or deny it myself. I can't say for certain that's what the stack for LNL would look like

SiliconFly · Apr 8, 2024

Ghostsonplanets said:
...alongside others power gating and limited power usage improvements...

Thats what they said about MTL too (and it didn't turn out well). Without seeing LNL in action, we can only hope it's better.

SiliconFly · Apr 8, 2024

coercitiv said:
I think this has been said repeatedly before, but it bears repeating: no single product matters anymore, only execution/iteration speed. LNL can be as mediocre as they come, the important part is how fast they're able to introduce the next meaningful upgrade.

Actually, only part of that is true. If LNC underperforms, Intel may never recover as lots rides on LNC. In it's current state, Intel is not strong enough to handle another blow of such magnitude.

poke01 said:
If Intel fails with LNL then I got no hope for them for the low power SoCs from them. They need to match or exceed M3 in ST(GB6) because optics.

I don't think LNL is even close to matching M3 in ST. It doesn't have to. It just needs to be just good enough.

FlameTail · Apr 8, 2024

SiliconFly said:
I don't think LNL is even close to matching M3 in ST. It doesn't have to. It just needs to be just good enough.

That sounds like copium.

In my opinion, if Lunar Lake can't atleast match the ST of X Elite, it's a failure.

dullard · Apr 8, 2024

FlameTail said:
That sounds like copium.

Lunar Lake should have a drastically improved NPU and GPU compared to Meteor Lake. It also will be down to ~8 W of power for fanless applications. Put that all together and it doesn't leave much power left for the CPU. Lunar lake's CPU is only going to be good enough for the NPU and GPU to do their work.

Lunar Lake is more for applications like ultralight laptops, tablets, handheld gaming, industrial controls, and commercial kiosks. It won't be a CPU workstation powerhouse.

SpudLobby · Apr 8, 2024

uzzi38 said:
I mean that's what the other forum dweller here has said but I can't confirm or deny it myself. I can't say for certain that's what the stack for LNL would look like

Thanks. That’s all I wanted. Agreed it’s all speculation.

SpudLobby · Apr 8, 2024

FlameTail said:
That sounds like copium.

In my opinion, if Lunar Lake can't atleast match the ST of X Elite, it's a failure.

Eh, it’s not copium because the M2/M3 is Apple and only competes to a very limited extent. A lot of people I know who switched (yes, Doug, it’s a small segment but it is real and that’s part of why LNL exists for Intel) wouldn’t have if Windows laptops which offer more variety and lower pricepoints iso-RAM/SSD had something tolerable.

Is LNL going to be mediocre if not laughable given it’s on N3B with 140mm^2 and it’s been 4 years since the M1, and Intel is the legacy chipmaker with resources? Absolutely IMO.

But given what I said about Apple, AMD isn’t even targeting the fanless or 17-30W *ultra efficient/long battery life* style chip market, and then the fact that not all of the major X86 Windows ecosystem is yet ported which hurts Qualcomm a bit*, and that Intel will have plenty of laptop design wins — Intel can sadly get away with something just ok relative to the year and the time/resources or the competition.

Doesn’t mean it’ll sell as well as it could without Qualcomm’s X Plus/Elite in the counterfactual, or without Strix. So they’re on backfoot. But like realistically it will ship a bit.

*For example and to my point:

LNL per the leaked documents will only match the M1 GPU perf @ 12W, not beat it, and it can beat the M2 albeit at higher power. So that’s just good enough, but not impressive in light of the time and area/node.

Yet even tho Qualcomm’s iGPU will be more efficient and same the CPU, Lunar Lake won’t have the games compatibility issue or emulation hit. So for someone wanting to play A) light games on a 15-20W ultrabook and otherwise use it for school/business, hard to recommend Snapdragon and Strix is not going to quite have the same target.

igor_kavinski · Apr 8, 2024

dullard said:
handheld gaming

It's gonna need some serious game engine patches to make sure that two of the least intensive threads fall on the E-cores. Otherwise, any time the main game engine thread gets scheduled on the E-core, there will be performance loss and/or hitching. Games could be a stuttering mess, especially those that depend heavily on multithreading. Four P-cores is less than the minimum six cores that most game engines have come to expect from gaming machines. There's a reason that AMD didn't introduce anything lower than the core count of Ryzen Z1. Lower core count would be detrimental to typical games.

dullard · Apr 8, 2024

igor_kavinski said:
It's gonna need some serious game engine patches to make sure that two of the least intensive threads fall on the E-cores. Otherwise, any time the main game engine thread gets scheduled on the E-core, there will be performance loss and/or hitching. Games could be a stuttering mess, especially those that depend heavily on multithreading. Four P-cores is less than the minimum six cores that most game engines have come to expect from gaming machines. There's a reason that AMD didn't introduce anything lower than the core count of Ryzen Z1. Lower core count would be detrimental to typical games.

Yes, heavily multithreaded games might not do well.

The ability for programmers to specify only P-cores for specific threads has been around since Alder Lake in 2021. If the game developers haven't added the one line of code to specify which threads shouldn't go on E-cores by the time Lunar Lake is out, that is a really sad failure on the game developers part.

Philste · Apr 8, 2024

From current Keynote, dude on Twitter posted it, although it's not openly available and probably under NDA.

Ghostsonplanets · Apr 8, 2024

dullard said:
Lunar Lake should have a drastically improved NPU and GPU compared to Meteor Lake. It also will be down to ~8 W of power for fanless applications. Put that all together and it doesn't leave much power left for the CPU. Lunar lake's CPU is only going to be good enough for the NPU and GPU to do their work.

Lunar Lake is more for applications like ultralight laptops, tablets, handheld gaming, industrial controls, and commercial kiosks. It won't be a CPU workstation powerhouse.

I'd argue 8W is plenty of power for GPU + NPU + CPU. Specially if you want your design to be on fanless models. Specially as we have Smartphones, Macbooks Air, SteamDeck, Switch, etc all being designs that sip power while providing high performance for their applications.

The problem lies at Intel inability at design P Cores that are efficient from the low to the high end power curve. That's something AMD is much better and they need it to be fixed with Lion Cove for LNL to shine.

SpudLobby said:
Intel will have plenty of laptop design wins

There wad a rumor last year that LNL has 3x as many design wins as MTL had. Intel claimed MTL had 200+ design wins at launch, so we're looking at LNL having 600+ design wins.

If it's true, it's a pretty rosy picture for OEMs confidence on LNL.

Ghostsonplanets · Apr 8, 2024

Philste said:
From current Keynote, dude on Twitter posted it, although it's not openly available and probably under NDA.
View attachment 96571 View attachment 96572 View attachment 96573

>100 peak TOPs??? So the BMG iGPU adds the XMX cores that MTL had removed? + the NPU is much more performant.

Also, first time I'm hearing Panther Lake is 18A, I3 and External. I assume External is the Celestial iGPU tile. 18A for Cougar Cove and Intel 3 for ??? (SoC?)

Tigerick · Apr 8, 2024

Ghostsonplanets said:
>100 peak TOPs??? So the BMG iGPU adds the XMX cores that MTL had removed? + the NPU is much more performant.

Also, first time I'm hearing Panther Lake is 18A, I3 and External. I assume External is the Celestial iGPU tile. 18A for Cougar Cove and Intel 3 for ??? (SoC?)

Panther Lake is new design with 2 tiles only. 18A is for CPU+SoC, tGPU is available for 2 version: Intel 3 for 4 XE and N3E for 12 XE

SiliconFly · Apr 8, 2024

Philste said:
From current Keynote, dude on Twitter posted it, although it's not openly available and probably under N

Intel always has the best slides in the market!

Tigerick said:
Panther Lake is new design with 2 tiles only. 18A is for CPU+SoC, tGPU is available for 2 version: Intel 3 for 4 XE and N3E for 12 XE

In the first slide, for the first time, Arrow Lake is being associated with Intel 3! Any ideas?

Tigerick · Apr 8, 2024

SiliconFly said:
Intel always has the best slides in the market!

In the first slide, for the first time, Arrow Lake is being associated with Intel 3! Any ideas?

Should be Arrow Lake-U with Intel 3's tCPU, replacing MTL-U

Philste · Apr 8, 2024

Ghostsonplanets said:
>100 peak TOPs??? So the BMG iGPU adds the XMX cores that MTL had removed? + the NPU is much more performant.

Bionic Squash already leaked that ARLs iGPU (also with XMX) alone has more than double the TOPs of whole Meteor Lake Combined (70+). LNLs iGPU probably isn't worse and then it has a 40+ TOPS NPU.

SiliconFly · Apr 8, 2024

Philste said:
View attachment 96571

The first slide says Intel 4 has been in production since H2 2023. That makes it a ELV node.

(Extremely Low Volume)

dullard · Apr 8, 2024

SiliconFly said:
In the first slide, for the first time, Arrow Lake is being associated with Intel 3! Any ideas?

There was a brief discussion in December 2021 from @IntelUser2000 and then @Ajay that Arrow Lake was on Intel 3. But no real details were given. http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=thread...ure-lakes-rapids-thread.2509080/post-40656166

SiliconFly · Apr 8, 2024

Tigerick said:
Should be Arrow Lake-U with Intel 3's tCPU, replacing MTL-U

Maybe. It's very confusing. LNC is a node agnostic design. But having it in 3 different nodes 20A & N3B & Intel 3 at the same time is just too ridiculous. It is either a very bad or a very desperate or a brilliant business decision. Not sure which.

SiliconFly · Apr 8, 2024

Tigerick said:
...Intel 3 for 4 XE...

Are you sure Intel 3 is for 4 Xe? Looks like it's actually for the tCPU after all. Not sure though.

Philste · Apr 8, 2024

SiliconFly said:
In the first slide, for the first time, Arrow Lake is being associated with Intel 3! Any ideas?

Is far is I know ARL-U is a Refresh of MTL om Intel 3. So basically ARL-U isn't ARL.

Hulk · Apr 8, 2024

I'm not hyped on NPU's, should I be? If I'm feeling lazy and I need a numerical solution to a math problem with no closed for solution or some other work that is boring for me to do manually I just hit up ChatGPT. It doesn't happen that often actually.

Am I stuck in the past and missing AI opportunities in my daily life/work?

Saylick · Apr 8, 2024

FYI - C&C has a new article on MTL's iGPU:

Intel’s Ambitious Meteor Lake iGPU

Intel and AMD both tried to ship iGPUs fast enough to compete with low end discrete cards over the past 10 years with mixed results. Recently though, powerful iGPUs have been thrown back into the s…

chipsandcheese.com

dullard · Apr 8, 2024

Hulk said:
I'm not hyped on NPU's, should I be? If I'm feeling lazy and I need a numerical solution to a math problem with no closed for solution or some other work that is boring for me to do manually I just hit up ChatGPT. It doesn't happen that often actually.

Am I stuck in the past and missing AI opportunities in my daily life/work?

What have you tried?

I like the potential for AI. Software like Teams can already summarize a meeting that you didn't attend and tell you what are your action items. Or have Outlook tell you what your boss things of topic X. Or have Copilot read a patent and explain how it might impact the design of what you are working on. Or have Excel tell you how data in column A impacts column B. All of these things already work in AI. And I can certainly understand why you don't want things like that done online with ChatGPT but would instead want it done locally on an NPU--especially when it deals with business secrets.

I think mostly we need training on what it can do now and how to use it effectively.

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

Senior member

Attachments

Senior member

Platinum Member

Golden Member

Golden Member

Platinum Member

Elite Member

Senior member

Senior member

Lifer

Elite Member

Member

Senior member

Senior member

Senior member

Golden Member

Senior member

Member

Golden Member

Elite Member

Golden Member

Golden Member

Member

Diamond Member

Diamond Member

Elite Member