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 (20A)	Arrow Lake (N3B)	Lunar Lake	Panther Lake
Platform	Mobile H/U Only	Desktop Only	Desktop & Mobile H&HX	Mobile U Only	Mobile H
Process Node	Intel 4	Intel 20A	TSMC N3B	TSMC N3B	Intel 18A
Date	Q4 2023	Q1 2025 ?	Desktop-Q4-2024 H&HX-Q1-2025	Q4 2024	Q1 2026 ?
Full Die	6P + 8P	6P + 8E ?	8P + 16E	4P + 4E	4P + 8E
LLC	24 MB	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.)

 

[BorisTheBlade82]

@Glo.
Says who? There are quite a lot of workloads that simply do not care about the GPU - SW development, Office, Rendering, just to name a few.

 

[Glo.]

@Glo.
Says who? There are quite a lot of workloads that simply do not care about the GPU - SW development, Office, Rendering, just to name a few.

For those workloads you have "standard" CPUs, and you are welcome to use them.

 

[poke01]

For APUs its not the CPU side that is the key, but its always the iGPU.

Till this day, we've always had an APU/SOC that always had CPU more powerful than the iGPU, but we're entering an era, where its the iGPU that is more powerful than the CPU.

And the performance of the iGPU will always define who won this round of competition. Not the CPU.

Well AMD and Intel are behind in this in the laptop space. Apple already created what you said with the M1 Max. As Anandtech notes and it's paraphrased "The M1 Max is more of a GPU than a CPU".

AMD may create it next year and Intel is going to start with Meteor Lake.

 

[TESKATLIPOKA]

Well AMD and Intel are behind in this in the laptop space. Apple already created what you said with the M1 Max. As Anandtech notes and it's paraphrased "The M1 Max is more of a GPU than a CPU".

AMD may create it next year and Intel is going to start with Meteor Lake.

Meteor Lake won't have a big IGP. There were some rumors about a 320EU IGP, but I am not sure for which chip, maybe Lunar Lake?

 

[Glo.]

Meteor Lake won't have a big IGP. There were some rumors about a 320EU IGP, but I am not sure for which chip, maybe Lunar Lake?

Arrow Lake-P, and later it was upgraded, by I think linux drivers, to 384 EUs.

Thats pretty big, if you ask me .

Meteor Lake-P on the other hand was downgraded from 192 EUs to 128, I think.

 

[IntelUser2000]

Actually, It looks like It was at ~2.1GHz. At 3GHz you can still shave off ~1/3 of power or gain ~15% of performance.

The iso power clock gain part is very optimistic at the ranges Intel plays at. Power will be better but still not as good as indicated. The N series chips based on the E cores clock far lower so they should benefit quite nicely.

 

[BorisTheBlade82]

For those workloads you have "standard" CPUs, and you are welcome to use them.

Quite the opposite - APUs are for people that do not care too much about GPU performance. Anyone else is welcome to use a dGPU. The majority of the sub 25w Notebook market couldn't care less about their APU bring able to run Crysis.
Apple is a bit special because at the moment they don't have dGPUs.

 

[poke01]

Quite the opposite - APUs are for people that do not care too much about GPU performance. Anyone else is welcome to use a dGPU. The majority of the sub 25w Notebook market couldn't care less about their APU bring able to run Crysis.
Apple is a bit special because at the moment they don't have dGPUs.

I don't think so. Why can't AMD make something like the Xbox Series X or PS5 APU but for laptops?

 

[BorisTheBlade82]

I don't think so. Why can't AMD make something like the Xbox Series X or PS5 APU but for laptops?

Because there is next to no market for that. Everything else being equal this would increase the price of a Notebook by what, maybe 200USD. If 19 out of 20 Notebooks are sold to people doing Office/Web/programming and the 20th could simply take a dGPU, then what is the point?
Back in the Fusion era I also thought that this would be the way to go - but GPGPU did not gain much traction and apart from Gamers simple iGPUs for Windows Desktops + Video acceleration seems to be sufficient.

 

[TESKATLIPOKA]

Because there is next to no market for that. Everything else being equal this would increase the price of a Notebook by what, maybe 200USD. If 19 out of 20 Notebooks are sold to people doing Office/Web/programming and the 20th could simply take a dGPU, then what is the point?
Back in the Fusion era I also thought that this would be the way to go - but GPGPU did not gain much traction and apart from Gamers simple iGPUs for Windows Desktops + Video acceleration seems to be sufficient.

True. Even for gaming, It doesn't matter how fast the IGP is, people(BFU) still think a separate gpu is a better option.
The problem with console APUs is that they use a lot of GDDR6 and that is costly.

I just remembered Intel's attempt at having a good cpu and gpu on a single package.
Intel made Hades Canyon which was a Kaby Lake+Vega on one package. It was pretty powerful at the time, but It wasn't very popular. Not sure If Intel didn't allow OEMs to put It inside laptops, but It was available only for Intel NUC.
i7-8809G manages 23 FPS in Cyberpunk 2077 at FullHD Ultra settings, Link. That's pretty good for something released in 2018.

 

[BorisTheBlade82]

@TESKATLIPOKA
IIRC there were some Dell Notebooks with it as well.

 

[Glo.]

Quite the opposite - APUs are for people that do not care too much about GPU performance. Anyone else is welcome to use a dGPU. The majority of the sub 25w Notebook market couldn't care less about their APU bring able to run Crysis.
Apple is a bit special because at the moment they don't have dGPUs.

The point for APUs is to move software towards "console" like apps, or 'iPad" apps.

There will be huge push from Microsoft for this type of sandboxation and launching of apps with next generation of their operating systems, and Linux is moving towards this for quite some time. Apple is already there. And this push, will require specific type of hardware. Which is not yet available. You need powerful GPU accelerators integrated into the CPU packages, for next generation of software.

 

[lightisgood]

ARL-S might be Intel_4 + TSMC_N3.
If ARL-S comes in early 2024, this is reasonable decision.

Intel Desktop CPU Rumors: Raptor Lake Refresh In 2023, Meteor Lake Up To Core i7 With 6+16, Arrow Lake Up To Core i9 With 8+16 SKUs In 2024

New rumors regarding Intel's next-gen desktop CPU lineup including Raptor Lake Refresh, Meteor Lake & Arrow Lake have been disclosed.

wccftech.com

> The Intel 15th Gen Arrow Lake-S Desktop CPUs will bring back the 24 cores that we get on Raptor Lake CPUs today.
> The Arrow Lake-S top die will utilize up to 24 cores which will be a combination of 8 Performance Cores and 16 Efficiency Cores.
> According to the Enthusiast Citizen, the lineup will only come in Core i7 and Core i9 flavors.
> The CPU will retain the Intel 4 (CPU) + TSMC N3 (GPU) SKU node layout.
> It is rumored that the 20A node won't make its way to the desktop lineup.

 

[BorisTheBlade82]

@Glo.
To be honest, what you describe is rather vague. I cannot see which kind of apps you have on your mind that would not be easily doable with today's APU of Intel and AMD.

 

[mikk]

So we have three candidates for ARL-S CPU tile: Intel 4, TSMC 3nm, Intel 20A

What is next? Maybe Intel 3 or TSMC 4nm? If there really is a Raptor Lake Refresh next year it clearly suggests MTL-S being late or not coming late 2023/early 2024 or won't replace i7/i9 Raptor Lake lineup.

 

[Hulk]

Meteor lake using Intel 4 process. Link or Link
Intel 4 process info where they mention -40% and +20%. Link

I also found this. Link

Actually, It looks like It was at ~2.1GHz. At 3GHz you can still shave off ~1/3 of power or gain ~15% of performance.

Thanks for the links.

MTL is starting to feel a bit like Broadwell in that with the node shrink issues Intel has had with both 14nm and 10nm I'm very suspicious that they will be able to maintain current Raptor Lake clocks. If 6+16 is true while the 16 may be able to surpass Gracemont due to the low current clocks of Gracemont and IPC improvements, I don't see them reaching 5.5 nT and 5.8 ST on the P's with MTL on Intel 4 with the first crack at this process.

They will then be in the "Broadwell dilemma." Release a desktop part with equal or worse desktop performance than the previous generation, which they mitigated somewhat with the huge cache on Broadwell or keep the release primarily mobile.

Mobile MTL would make a lot of sense considering the power savings projected above. Mobile parts today are primarily power/thermally constrained so at these sub 5GHz, which is where mobile lives a 20% gain in performance and/or 40% gain in power would be a huge step forward.

After MTL Intel would probably have Intel 4 refined enough to hit Raptor-like clocks in the same way clocks steadily increased during the Skylake era.

 

[Hulk]

Well AMD and Intel are behind in this in the laptop space. Apple already created what you said with the M1 Max. As Anandtech notes and it's paraphrased "The M1 Max is more of a GPU than a CPU".

AMD may create it next year and Intel is going to start with Meteor Lake.

Depends on how you define "behind." Generally the important metric for large (and small) corporations is sales and in that regard they (especially Intel) are not behind.

 

[poke01]

Depends on how you define "behind." Generally the important metric for large (and small) corporations is sales and in that regard they (especially Intel) are not behind.

I meant in terms in implementation of the tech.

 

[Geddagod]

So we have three candidates for ARL-S CPU tile: Intel 4, TSMC 3nm, Intel 20A

What is next? Maybe Intel 3 or TSMC 4nm? If there really is a Raptor Lake Refresh next year it clearly suggests MTL-S being late or not coming late 2023/early 2024 or won't replace i7/i9 Raptor Lake lineup.

I don't think Intel 4 was ever a runner for ARL-S cpu tile. I have no idea where WWCFTech got that, the original post never said that.
I think the RPL refresh is just higher clocks. The post mentions that it just boosts 100 or 200 mhz faster and no higher core counts. This is just speculation on my part, but maybe the L3 cache per core increases as well, like it is rumored to do for emerald rapids.
I agree though, the only reason for a RPL refresh coming that late, if it is coming, is because MTL-S is not on track for a late Q3/Q4 2023 launch.

 

[igor_kavinski]

I agree though, the only reason for a RPL refresh coming that late, if it is coming, is because MTL-S is not on track for a late Q3/Q4 2023 launch.

It's funny. RPL itself is ADL+. RPL-R would be ADL++. Intel loves the plusses.

 

[Exist50]

It's funny. RPL itself is ADL+. RPL-R would be ADL++. Intel loves the plusses.

Skylake all over again.

 

[eek2121]

iGPUs have uses outside gaming. Windows has been gradually using the GPU for certain tasks as we have gone from Windows 7 through Windows 11.

EDIT: Virus scanning, for example.

I suspect we will see more usage as software and hardware continue to evolve.

EDIT 2: ...and here is a slide deck that discusses using the GPU for compression/decompression. NOT used in Windows currently, but possible: https://on-demand.gputechconf.com/g...-parallel-lossless-compression-using-gpus.pdf

 

[Geddagod]

It's funny. RPL itself is ADL+. RPL-R would be ADL++. Intel loves the plusses.

This might not be *technically* the most accurate, but just for the memes
Golden Cove (client), Golden Cove + (server with extra L2 cache, etc etc), Golden Cove ++ (RPL better prefetcher and extra L2 cache, better clocks), Golden Cove +++ ?(rumored RPL refresh 2h 2023, higher clocks), Golden Cove ++++ (Emerald Rapids, rumored to use Raptor Cove but with increased L3 cache per core), Golden Cove +++++ (MTL redwood cove), Golden Cove ++++++ ?(New redwood cove + core used in Granite Rapids)

 

[IntelUser2000]

What is next? Maybe Intel 3 or TSMC 4nm? If there really is a Raptor Lake Refresh next year it clearly suggests MTL-S being late or not coming late 2023/early 2024 or won't replace i7/i9 Raptor Lake lineup.

What? It won't. Arrowlake replaces Raptorlake for high end.

@Hulk Yea, except in this case "Skylake" comes as high end at the same time as Broadwell.

 

[BorisTheBlade82]

iGPUs have uses outside gaming. Windows has been gradually using the GPU for certain tasks as we have gone from Windows 7 through Windows 11.

EDIT: Virus scanning, for example.

I suspect we will see more usage as software and hardware continue to evolve.

EDIT 2: ...and here is a slide deck that discusses using the GPU for compression/decompression. NOT used in Windows currently, but possible: https://on-demand.gputechconf.com/g...-parallel-lossless-compression-using-gpus.pdf

That is exactly what I meant - these things are rather old.
We have GPU acceleration on the Desktop now for more than a decade. Back in the "Future is Fusion" days there was a lot of buzz about what you could do with integrating graphics and doing APUs. That was around 2010. But apart from Casual Gaming I see no killer applications that could make good use of much more powerful iGPUs than what we have today.