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.)

 

[uzzi38]

Could you please show us the other slides you mention?

Oh, I totally forgot to come back to this.

Here. I'm not leaking anything that's new, so I've censored out anything that has even the slightest bit of being actually new information. Deal with it.

 

[inf64]

That is pretty abysmal ST gain for Arrow lake, I find it very hard to believe it. The only way this could be true is that Arrow Lake clocks waaay lower than Raptor Lake and this nullifies the IPC gains.

 

[S'renne]

That is pretty abysmal ST gain for Arrow lake, I find it very hard to believe it. The only way this could be true is that Arrow Lake clocks waaay lower than Raptor Lake and this nullifies the IPC gains.

Probably if Intel decides to focus on mainly efficiency over absolute performance due to inherent design flaws I think

 

[SiliconFly]

Oh, I totally forgot to come back to this.

Here. I'm not leaking anything that's new, so I've censored out anything that has even the slightest bit of being actually new information. Deal with it.

View attachment 85836

When I last mentioned that ARL may get Battlemage as it's already up & running, people said no. Many believe ARL is actually getting Alchemist+. But this slide says otherwise.

So, is this slide legit? Is ARL getting Battlemage?

 

[uzzi38]

When I last mentioned that ARL may get Battlemage as it's already up & running, people said no. Many believe ARL is actually getting Alchemist+. But this slide says otherwise.

So, is this slide legit? Is ARL getting Battlemage?

It very clearly says discrete GPU, not integrated.

 

[dullard]

Oh, I totally forgot to come back to this.

Here. I'm not leaking anything that's new, so I've censored out anything that has even the slightest bit of being actually new information. Deal with it.

Thanks for sharing. I have not seen that before. So you were talking about gaming performance. That helps clarify. Although it doesn't specify if that "perf gain" refers to IPC, clocks, or the combination of both.

 

[inf64]

Thanks for sharing. I have not seen that before. So you were talking about gaming performance. That helps clarify. Although it doesn't specify if that "perf gain" refers to IPC, clocks, or the combination of both.

Well ST gains must be a cumulative effect of clocks and IPC, or else they would specify IPC (or clock). Historically, intel used ST uplift this way, correct me if I'm wrong.

 

[JoeRambo]

Arrow-Lake-S perf seems to be spot on where earlier leaks ( internal perf estimate? ) predicted.

I am not sure what people expect from more iCache, 3MB of L3 and similar L3 cache setup? It's same core, but enabled to get better MT performance.
Absolutely nothing radical that would increase perf in unexpected way like 100MB of L3 or completely new design of P core -> that will have to wait for redesign.

I think the people who expect wonders from Zen5 might be also surprised in performance increase that is smaller than expectation in first generation of new core.

 

[inf64]

Arrow-Lake-S perf seems to be spot on where earlier leaks ( internal perf estimate? ) predicted.

I am not sure what people expect from more iCache, 3MB of L3 and similar L3 cache setup? It's same core, but enabled to get better MT performance.
Absolutely nothing radical that would increase perf in unexpected way like 100MB of L3 or completely new design of P core -> that will have to wait for redesign.

I think the people who expect wonders from Zen5 might be also surprised in performance increase that is smaller than expectation in first generation of new core.

Although I agree with almost everything you said, I think you are wrong regarding Zen 5. Zen 5 might disappoint when it comes to peak GHz target, but IPC jump should be higher than what Zen 3 brought - and this would put it in the Bulldozer -> Zen 1 category of more than 20% IPC jump between two gens.

 

[ondma]

Arrow-Lake-S perf seems to be spot on where earlier leaks ( internal perf estimate? ) predicted.

I am not sure what people expect from more iCache, 3MB of L3 and similar L3 cache setup? It's same core, but enabled to get better MT performance.
Absolutely nothing radical that would increase perf in unexpected way like 100MB of L3 or completely new design of P core -> that will have to wait for redesign.

I think the people who expect wonders from Zen5 might be also surprised in performance increase that is smaller than expectation in first generation of new core.

Why do you say that? Do you have any recent leaks?
I thought ARL was supposed to be the first itineration of the new core design? It doesn't make sense that they could not get hyperthreading working (if rumor is correct) if it was the same basic core as ALD/RL.

 

[Kepler_L2]

Arrow-Lake-S perf seems to be spot on where earlier leaks ( internal perf estimate? ) predicted.

I am not sure what people expect from more iCache, 3MB of L3 and similar L3 cache setup? It's same core, but enabled to get better MT performance.
Absolutely nothing radical that would increase perf in unexpected way like 100MB of L3 or completely new design of P core -> that will have to wait for redesign.

I think the people who expect wonders from Zen5 might be also surprised in performance increase that is smaller than expectation in first generation of new core.

It's a 30-50% wider core than GLC. A 5% improvement is extremely disappointing.

 

[inf64]

It's a 30-50% wider core than GLC. A 5% improvement is extremely disappointing.

The only way 5% figure could be true is if the new core clocks topped at roughly 4.8-5Ghz for ST.

 

[SiliconFly]

Why do you say that? Do you have any recent leaks?
I thought ARL was supposed to be the first itineration of the new core design? It doesn't make sense that they could not get hyperthreading working (if rumor is correct) if it was the same basic core as ALD/RL.

Yep. No hyper-threading means only one thing. It's LNC.

But the "no hyper-threading" thing is still a rumor.

 

[SiliconFly]

It's a 30-50% wider core than GLC. A 5% improvement is extremely disappointing.

If ARL-S isn't getting any noticeable performance improvement, then it's going to get a lot of improvement in power-efficiency. A lot!

Remember, ARL is in Intel 20A which has a massive PPW gain of 36% over Intel 4.

 

[A///]

Probably if Intel decides to focus on mainly efficiency over absolute performance due to inherent design flaws I think

And let AMD eat more of their lunch? That's a laughable thought.

 

[A///]

Do you have any recent leaks?

I imagine from his hind quarters. It's not the first time he's downplayed the reality that's coming in a few months.

 

[Geddagod]

I am not sure what people expect from more iCache, 3MB of L3 and similar L3 cache setup? It's same core, but enabled to get better MT performance

No.

or completely new design of P core

That's LNC. Though maybe not "completely" new design, it's just more the same of "wider, deeper, smarter"

thought ARL was supposed to be the first itineration of the new core design?

It's not Royal Core.

If ARL-S isn't getting any noticeable performance improvement, then it's going to get a lot of improvement in power-efficiency. A lot!

That's not necessarily true.

Remember, ARL is in Intel 20A which has a massive PPW gain of 36% over Intel 4.

That's not how it works.

 

[A///]

Royal Core.

Much like who ate the last piece of pie, what's going on with this?

 

[SiliconFly]

That's not how it works.

For the same architecture, it does. For a better architecture, even more.

 

[H433x0n]

Oh, I totally forgot to come back to this.

Here. I'm not leaking anything that's new, so I've censored out anything that has even the slightest bit of being actually new information. Deal with it.

View attachment 85836

is that 1.05 over RPL-R? That puts it at +10% over RPL which still isn’t great but it’s better than nothing.

 

[SiliconFly]

is that 1.05 over RPL-R? That puts it at +10% over RPL which still isn’t great but it’s better than nothing.

Hard to say whether it's RPL or RPL-R. Absolutely no context. We shouldn't be using this info until we have better clarity.

 

[Glo.]

And let AMD eat more of their lunch? That's a laughable thought.

Mobile is where the battle will be the biggest and toughest between AMD and Intel in upcoming years.

 

[dullard]

Remember, ARL is in Intel 20A which has a massive PPW gain of 36% over Intel 4.

Do note: that your math is on the high end of what to expect.

Intel 3 is supposedly ~18% more performance per watt compared to Intel 4 (everything else being equal)
Intel 20A is UP TO 15% more performance per watt compared to Intel 3 (everything else being equal)

So crude math would be 1.18 * 1.15 = 1.357 or 35.7% more performance per watt (everything else being equal). So right off the bat you rounded up. But also you ignored the UP TO part. So, think of 35.7% as one of the better case scenarios (everything else being equal).

But of course, not everything is equal, so Intel could ramp up the frequency. Or Intel could add other features that use energy. Etc. All of those would toss performance per watt out the window. That and we still have the rumor that not all Arrow Lake is on 20A.

 

[SiliconFly]

Mobile is where the battle will be the biggest and toughest between AMD and Intel in upcoming years.

If Intel's hyper-focused on efficiency, then ARL mobile's power-efficiency will be off the charts due to 20A. Will definitely have a huge impact in mobile.

ARL-S (desktop) can also have a huge PPW improvement. But without much performance improvement how that translates into sales is a big if.

 

[adroc_thurston]

will also have a huge PPW improvement

15% nT perf ISO core count is bad and you know it.