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

[Tigerick]

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	12 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 (N3B)	Lunar Lake	Panther Lake
Platform	Mobile H/U Only	Desktop & Mobile H&HX	Mobile U Only	Mobile H
Process Node	Intel 4	TSMC N3B	TSMC N3B	Intel 18A
Date	Q4 2023	Desktop-Q4-2024 H&HX-Q1-2025	Q4 2024	Q1 2026 ?
Full Die	6P + 8P	8P + 16E	4P + 4E	4P + 8E
LLC	24 MB	36 MB ?	12 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.)

 

[511]

xtor density metric doesn't really matter. You can't even compare them ISO anyway. Plus it varies tremendously with cache sizes.

Only thing that matters is the performance delivered per mm2. In that case Intel still sucks for Lion Cove and Battlemage. But Skymont for example is pretty damn good.

Directly or indirectly these factors are all related for a node.
HP Cell -> low xtor density -> lower perf/mm2 -> higher perf/xtor
HD Cell -> higher xtor density -> higher perf/mm2 -> lower perf/xtor

Which would really be weird since X3D is about gaming. What "ST" gaming is there nowadays?

Yeah it's basically has become a cache competition.

Edit: Fixed Typo

 

[511]

And when was Intel the consensus leader? In 2017 when they had the single thread and gaming crown. For 2026 how about AMD 12C/24T on a single CCD, with v-cache for the gaming champ? All this actually begs the question though, I seriously doubt Intel will actually produce the 16 + 32 core variant, with their limited resources and renewed interest in increasing margins. A 48 core will just be too expensive to produce and sell at a reasonable cost.

It is not tbh it's same strategy as AMD's Chippet design it entirely depends on the die size for their compute tile cause that is only external tile in NVL with IO.

 

[DavidC1]

Directly or indirectly these factors are all related for a node.
HP Cell -> low xtor density -> lower perf/mm2 -> higher perf/xtor
HD Cell -> higher xtor density -> lower perf/mm2 -> lower perf/xtor

You are right.

But we can't really determine xtor density at a chip level. It's almost impossible. The manufacturers use different metrics, and you can't compare between CPUs vs GPUs obviously. Perf/mm2, Perf/$, Perf/W, those are things we can actually test.

 

[DrMrLordX]

So nobody cared about the introduction of 1800X and the impact of that?

The 1800x represented a massive improvement in desktop processing power versus AMD's previous gen Bulldozer chips (not to speak of Steamroller and Excavator, which were too limited in scope). ST and MT performance went through the roof. In contrast, the 7700k was sort of the lame duck successor the the 6700k (due to 14nm delays).

Unless Nova Lake-S shows major ST and MT uplifts over Arrow Lake-S, it probably won't turn many heads. Flagship Zen6 will be a nice improvement over the 9950X and will definitely not be a warmed-over rehash product like the 7700k!

But who knows, maybe what people really want is a Cinebench accelerator.

 

[DavidC1]

But who knows, maybe what people really want is a Cinebench accelerator.

Which would be fine, if it compares well with AMD, their primary competitor. Super fast in embarassingly multi-threaded workloads, beats competition slightly in ST and gaming, and uses about the same power.

Sure, it wouldn't turn heads, but certainly a turnaround from today.

The 1800x represented a massive improvement in desktop processing power versus AMD's previous gen Bulldozer chips (not to speak of Steamroller and Excavator, which were too limited in scope).

AMD basically gave you Intel's expensive HEDT at a much lower price, while catching up dangerously close in less threaded applications. If 1800X wasn't that enormously faster than the 7700K in MT, then it wouldn't have been looked upon so fondly. But it was.

 

[511]

Btw I wonder what's the disadvantage of KF SKU vs K SKUs cause you get the display and media engine with KF AND K SKUs you don't need iGPU for display from motherboard.

 

[511]

The 1800x represented a massive improvement in desktop processing power versus AMD's previous gen Bulldozer chips (not to speak of Steamroller and Excavator, which were too limited in scope). ST and MT performance went through the roof. In contrast, the 7700k was sort of the lame duck successor the the 6700k (due to 14nm delays).

Unless Nova Lake-S shows major ST and MT uplifts over Arrow Lake-S, it probably won't turn many heads. Flagship Zen6 will be a nice improvement over the 9950X and will definitely not be a warmed-over rehash product like the 7700k!

But who knows, maybe what people really want is a Cinebench accelerator.

They went the way of removing HT If it doesn't achieve 10% ST improvement over Zen 6 than the P core team has to DIP themselves in a glass full of water without APX.

 

[AcrosTinus]

Hopefully people will fondly remember Intel's push for more MT performance while keeping ST perf competitive. I really hope that being competitive at gaming is not morphed into "is bad for gaming" or cannot game....

 

[Fjodor2001]

AMD basically gave you Intel's expensive HEDT at a much lower price

Yes, just like 54C NVL-S will give you expensive HEDT such as ThreadRipper at a much lower price. At least up to ~32C level of TR performance.

 

[Josh128]

Yes, just like 54C NVL-S will give you expensive HEDT such as ThreadRipper at a much lower price. At least up to ~32C level of TR performance.

mmm Im skeptical of it matching 32C of Zen 5 TR in MT. TR has 350W+ of TDP to play with, basically double the power of a 9950X, which itself is 200W (170W TDP) of electrical power at stock. I mean, maybe, if its really 18A and they can pull some impressive perf/w efficiency out of that, but is Intel really that hellbent on dying on the 300W+ on desktop hill?

 

[vanplayer]

All this actually begs the question though, I seriously doubt Intel will actually produce the 16 + 32 core variant, with their limited resources and renewed interest in increasing margins. A 48 core will just be too expensive to produce and sell at a reasonable cost.

Well, cuz it's just an old rumor.

It seems that even some Intel employees don't know what would NVL-S look like when launch.

 

[DrMrLordX]

If 1800X wasn't that enormously faster than the 7700K in MT, then it wouldn't have been looked upon so fondly. But it was.

My main point was that if the 1800X hadn't been enormously faster than the FX chips it wouldn't have been looked upon so fondly.

 

[Thunder 57]

Yes, just like 54C NVL-S will give you expensive HEDT such as ThreadRipper at a much lower price. At least up to ~32C level of TR performance.

If it performs similarly why would they release it at a much lower price?

 

[Fjodor2001]

If it performs similarly why would they release it at a much lower price?

For the same reasons as for 1800X

 

[Thunder 57]

For the same reasons as for 1800X

Not comparable. AMD's market share was crap back then. Intel still has the market share lead. Also, The 1800X was still behind in ST performance.

 

[DavidC1]

Not comparable. AMD's market share was crap back then. Intel still has the market share lead. Also, The 1800X was still behind in ST performance.

If you look at their revenue the company is being held up almost entirely by the client revenue not dropping much. Once they lose that, you'll see the revenue fall drastically.

AMD back then was also much leaner. Intel is not like that.

but is Intel really that hellbent on dying on the 300W+ on desktop hill?

They might have to. Surely new processes with fraction of the older gains isn't going to win you much. Also it's possible same is happening on the design side and that will flatline.

 

[Thunder 57]

If you look at their revenue the company is being held up almost entirely by the client revenue not dropping much. Once they lose that, you'll see the revenue fall drastically.

AMD back then was also much leaner. Intel is not like that.

They might have to. Surely new processes with fraction of the older gains isn't going to win you much. Also it's possible same is happening on the design side and that will flatline.

So no fire sales other than the occasional Core Ultra 7 265k?

 

[511]

mmm Im skeptical of it matching 32C of Zen 5 TR in MT. TR has 350W+ of TDP to play with, basically double the power of a 9950X, which itself is 200W (170W TDP) of electrical power at stock. I mean, maybe, if its really 18A and they can pull some impressive perf/w efficiency out of that, but is Intel really that hellbent on dying on the 300W+ on desktop hill?

NVL has ~1.5 Node shrink and better cores as well i would not be surprised at 2X MT perf vs 285K.

 

[igor_kavinski]

i would not be surprised at 2X MT perf vs 285K.

Maybe in cinememe.

If it manages 2x MT perf in bandwidth limited scenarios, Intel stock price will shoot through the roof.

 

[jpiniero]

If you look at their revenue the company is being held up almost entirely by the client revenue not dropping much. Once they lose that, you'll see the revenue fall drastically.

Raptor Lake 4eva

 

[Io Magnesso]

In the first place, in a consumer environment, I think that 16 cores or more is not enough bandwidth…
Whether it is Intel or AMD,

 

[Io Magnesso]

No matter how much 52 cores are installed, it is unlikely that the TDP will exceed 300W by default...
Depends on the power settings preset But

 

[msj10]

Maybe in cinememe.

If it manages 2x MT perf in bandwidth limited scenarios, Intel stock price will shoot through the roof.

The bLLC variant will have 180MB L3 so that should help with bandwidth.

 

[eek2121]

Havent been following CPUs much lately, due to Intel not being very competitive. What is the consensus: Is NL going to really double the core counts, or is this just another rumor/hope that will ultimately get cancelled like so many other projects from Intel lately? It sounds great, but upon thinking about it, what would this accomplish? How would they manage the heat and power usage, unless the new core design is hugely more efficient. I know they will be on a more efficient process, but that couldnt possibly save enough power to double the core counts. The problem with ARL seems to be latency, not raw processing power anyway. If they do this, and can get the power and temps under control, it would obviously be a multi threading beast, but for gaming??

Core counts will only go up on specific products, and only because it is dead simple for them to add a second chiplet.

More importantly, is Intel gonna be here in 10 years? Is x86 gonna be a thing in 10 years? Very clouded the future is.

x86 isn’t going anywhere.

NVL-S 16+32 is 320-400W IIRC, which will probably still be power limited with so many cores, but enough to beat 24 core Zen6 in Cinebench which I assume is the goal.

When you count hyperthreading, they will be pretty close in terms of performance. Power limits and IPC will be the deciding factor.

LMAO this is a Workstation Class PL I thought it was ~300W a classic case of Intel being Intel.
I don't know how much turbo they want it to have.

bLLC exis for this comparison 🙂.

If it performs similarly why would they release it at a much lower price?

They won’t. It isn’t even confirmed to be a Core X part. It could end up being an Xeon workstation part.

If it ends up being a Core 9 or whatever Intel calls their stuff these days, it will probably be $750-$800

 

[LightningZ71]

Somewhat faster vs what? Zen 5 vanilla or Arrow Lake?

Vanilla Zen5. It should make up ground on Arrow Lake as well, but, it's going to be very game dependent with respect to how sensitive the individual game is to L3 cache size. All of this notwithstanding the actual core improvements in Zen6...