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	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)	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 ?	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.)

 

[TwistedAndy]

So the 4.7/4.0 clocks were with PL1 & PL2 = 250W? If so, do we really expect the retail clocks to be much higher? 250W is already rather high.

It's an engineering sample. The clocks and power consumption are not final and will be optimized.

If we get a 3% better score for Arrow Lake running on 4.7GHz than Raptor Lake on 5.4 - 5.8GHz, it's a great result (~15-20% ST performance increase)

 

[cebri1]

My guess is still at 8-10% ST improvement but those numbers look promising especially regarding MT because the e-cores should go up to 4.6 if I’m not mistaken.

 

[Wolverine2349]

I'm a little unclear on something regarding that tweet.

So the 4.7/4.0 clocks were with PL1 & PL2 = 250W? If so, do we really expect the retail clocks to be much higher? 250W is already rather high.

250 watts at 4.7GHz P and 4GHz e is already high power. Immpressive IPC uplift, but 300W plus for 5GHxz clocks on a tiny die and such heat density. Not good and will be impossible to cool on air or if so dump excess heat inside the case and be bad for RTX 4090 users in enclosed sound dampened cases.

Yikes. Depressing thought.

 

[Wolverine2349]

So if the 4.7Ghz sample is 3% ahead of a 5.7Ghz 13900k (correct me if im wrong but i think this is the 1t boost clock of it), that would be ~25% higher ipc on a ES2 sample. Couple with with a final step and some microcodes, the ipc boost could be a little higher. If this thing can clock like raptor lake, or at least 5.7 and 4.6 like the chinese leaker said the qs were clocking, it will be considerably faster than raptor lake

Can they do that at reasonable power consumption though so too much excess heat not dumped into the case? Like 200watts or lower or maximum 250 watts at 5.7GHz P-cores and 4.4 to 4.6GHz e-cores?

 

[Philste]

So the 4.7/4.0 clocks were with PL1 & PL2 = 250W? If so, do we really expect the retail clocks to be much higher? 250W is already rather high.

Well, it doesn't mean it took 250W, it could've used only 25W, who knows? It's only the upper limit at the end. Old i5 3350P in my parents Office PC is rated with 69W TDP and uses 38W in Prime95. I know many can't believe this, but there were times where Intel CPUs didn't reach their own specified TDP.

 

[poke01]

I’m certainly more optimistic now

 

[Wolverine2349]

Well, it doesn't mean it took 250W, it could've used only 25W, who knows? It's only the upper limit at the end. Old i5 3350P in my parents Office PC is rated with 69W TDP and uses 38W in Prime95. I know many can't believe this, but there were times where Intel CPUs didn't reach their own specified TDP.

Yeah forgot about. PL1 abd PL2 at 250w just means the CPU can use 250w maximum if need be but does not have too.

Hopefully it took way less.

Really hoping for 200w or less 25 percent IPC performance uplift and rock stability at that without easy fragile degradation.

Hopefully Raptor Lake i9 and i7s this ain't regarding stability and degradation issues.

 

[Goop_reformed]

Immpressive IPC uplift

But at what cost? 4.7ghz @250W is already insane, going above 5.0ghz would be ~ 350W for sure. This is insane

 

[poke01]

But at what cost? 4.7ghz @250W is already insane, going above 5.0ghz would be ~ 350W for sure. This is insane

It’s doesn’t mean it’s using 250watts the limit is 250w

 

[controlflow]

But at what cost? 4.7ghz @250W is already insane, going above 5.0ghz would be ~ 350W for sure. This is insane

Where does it say that it took 250W? You can set PL1 and PL2 to whatever, it doesn't necessarily mean those numbers are what was actually consumed, nor does it mean that it needs those power limits to hit those frequencies. Especially in an ES chip, the frequencies and power limits would not be optimized yet.

 

[Goop_reformed]

It’s doesn’t mean it’s using 250watts the limit is 250w

Didn't the leaker say it was tested at 250W or he said limit is 250. Either way fat/bloated cores tend to drink the wattage.

Where does it say that it took 250W? You can set PL1 and PL2 to whatever, it doesn't necessarily mean those numbers are what was actually consumed, nor does it mean that it needs those power limits to hit those frequencies. Especially in an ES chip, the frequencies and power limits would not be optimized yet.

Chip is basically done, a couple of months won't change anything.

 

[AcrosTinus]

Finished watching the Wendell from Level1techs investigation into the Intel crash issue.

The Problem might be even more serious than I anticipated.

 

[H433x0n]

I wouldn’t read too far into this latest revelation. I think JKin may have made a mistake.

The 250W PL2 doesn’t mean it was actually using 250W either - that was just the configured limit.

 

[Saylick]

I wouldn’t read too far into this latest revelation. I think JKin may have made a mistake.

The 250W PL2 doesn’t mean it was actually using 250W either - that was just the configured limit.

Yeah, it doesn't help that they deleted their tweet, too.

 

[H433x0n]

Yeah, it doesn't help that they deleted their tweet, too.

I’m in a discord server with him and I think he deleted it because he made a mistake.

ARL is still likely to just be 5-10% faster 1T over RPL.

 

[poke01]

Finished watching the Wendell from Level1techs investigation into the Intel crash issue, well damn....
If they don't get their messaging on point ASAP, this might hurt Arrow Lake badly.
They cannot pinpoint the source of the Issue as of now and are soon introducing a new mainstream CPU that is vastly more complex than the monolithic RPL-R.

Hmmm... Maybe I have to find my inner ZEN.

I’m already thinking of Zen 5. The 12400 is bottlenecking my 4070S.

I hope Intel addresses this, I never had a AMD CPU but I can’t trust them with next gen until this instability is fixed.

 

[Wolverine2349]

I’m already thinking of Zen 5. The 12400 is bottlenecking my 4070S.

I hope Intel addresses this, I never had a AMD CPU but I can’t trust them with next gen until this instability is fixed.

Trust who Intel or AMD? You mean Intel Raptor Lake CPU instability and would have a hard time trusting getting Arrow Lake due to Raptor Lake instability issues?

Or otherwise you may go AMD Zen 5.

I am on 7800X3D right now.

I really would like to see more than 8 cores on a single die/node/CCX-CCD/ring bus/tile with a homogenous arch to get the best of set and forget it CPU for games that can start to scale to more than 8 cores and bets set and forget it with no APO nor process Lasso for best compatibility with games last 15 years and future games too as set and forget.

No such CPU sadly appears to be in the cards form either AMD or Intel. AMD still has the dual CCX-CCD going more than 8 cores. Intel is heterogenous beyond 8 cores.

 

[poke01]

Trust who Intel or AMD? You mean Intel Raptor Lake CPU instability and would have a hard time trusting getting Arrow Lake due to Raptor Lake instability issues?

Yes I mean Intel. The only solace is that’s it’s ARL is not on Intel 7 and is on N3B.

But if it’s a architecture problem and not a node problem, hopefully they fixed it in ARL.

 

[Hitman928]

Yes I mean Intel. The only solace is that’s it’s ARL is not on Intel 7 and is on N3B.

But if it’s a architecture problem and not a node problem, hopefully they fixed it in ARL.

I don't think it's an architecture problem (obviously we don't know for sure yet) but it is seeming to me more and more that Intel just pushed their bins too hard to try and keep up with AMD and a significant amount of CPUs being sold with boosts beyond 5.5 GHz or so just can't handle it (multiple variables that go into it but I think this is what it is coming down to).

 

[Wolverine2349]

Yes I mean Intel. The only solace is that’s it’s ARL is not on Intel 7 and is on N3B.

But if it’s a architecture problem and not a node problem, hopefully they fixed it in ARL.

Yes very true. Though might be both? Who knows?

Though Alder Lake is also on 10nm node and it seems ok? Has your 12400 been stable?

I have heard mostly reports of problems with i7 and i9 Raptor Lake. Have not heard about i9 and i7 Alder Lake problems.

Seems like maybe a Raptor Lake arch problem or maybe the 8 + 16 die has issues on 10nm node unlike 8=8 Alder Lake die? Its a little smaller on same sized node to fit the 8 extra e-cores. Or maybe it is an arch problem or both.

The IMC with DDR5 is so bad and fragile on Raptor Lake any 4 DIMM board even with only 2 DIMMs XMP 6000+ is not truly ever stable even if you think it is. Run OCCT Large Data Set variable and you will see how unstable it is with a random WHEA or CPU core errors. Turn XMP off they are gone. XMP 6000+ works fine on non Asus 2 DIMM boards.

But Raptor Lake has other non IMC stability issues with like almost no fix like the random ring bus errors and stability issues and degradation is a real thing not fixed even by a non-Asus 2 DIMM board.

Arrow Lake better not have that issue or even the IMC one that requires a 2 DIMM board for true stability.

 

[poke01]

Has your 12400 been stable?

yes, the maximum is 85 watts even in Prime 95. It’s a nice chip but lacks in MT due to 6P cores.

 

[Det0x]

It appears that Jaykihn's results, which showed a 3% ST and 15% MT increase, were obtained on the 4.7/4.0 GHz clocks:

https://twitter.com/x/status/1811123054841299020

Obviously, the retail CPUs will use higher clocks.

New update again

 

[cebri1]

That guy seems pretty unreliable then if he is reporting performance estimations and doesn’t even know the clock speeds.

 

[DrMrLordX]

New update again

Huh, wonder what kind of boosting parameters it did have?

 

[inf64]

Huh, wonder what kind of boosting parameters it did have?

Well there were rumors of 5.6 or 5.7Ghz boost for ST, but not sure how reliable were they.