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)	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.)

 

[mikk]

LNL 64EU is like MTL 128EU since Xe2 has twice the ALU units per EU. LNL should beat MTL with ease at 15W because MTL struggles below 28W.

 

[dullard]

Is LNL's use case gaming?

Gaming is one use case. I think the biggest use cases would be industrial computers and kiosks (think any screen you interact with such as a help-desk, digital menu, information booth, point of sale systems, etc). There low power fanless applications are almost a must-have.

I'd really like them to use the AI features at stores at the end of aisles. I would love to be able ask where XYZ is located and have it say AISLE 18, lower shelf, bin #7.

 

[tamz_msc]

LNL 64EU is like MTL 128EU since Xe2 has twice the ALU units per EU. LNL should beat MTL with ease at 15W because MTL struggles below 28W.

And it has double the L2 at 8 MB vs 4 MB for MTL.

 

[Thunder 57]

Gaming is one use case. I think the biggest use cases would be industrial computers and kiosks (think any screen you interact with such as a help-desk, digital menu, information booth, point of sale systems, etc). There low power fanless applications are almost a must-have.

I'd really like them to use the AI features at stores at the end of aisles. I would love to be able ask where XYZ is located and have it say AISLE 18, lower shelf, bin #7.

Are we calling that "AI" now? Rathar than just looking something up?

 

[jpiniero]

Are we calling that "AI" now? Rathar than just looking something up?

Everything is AI now.

 

[Goop_reformed]

So is it good now? or my previous comment still stands

So in summary:

New core: check
New node: check
New graphics tile: check
New and expensive packaging tech: check

The thing barely competes with Phoenix which was release 3 quarters before. Next please.

 

[igor_kavinski]

Page 94 - Question - AMD Phoenix/Zen 4 APU Speculation and Discussion

Page 94 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Intel's world class engineering eked out a MASSIVE SIX EXTRA MINUTES of battery time from MTL H to U series.

Truly groundbreaking power efficiency!

Snapdragon Elite X stands no chance in challenging the reigning champion!

Humanity will rejoice! Global warming will be a thing of the past now! Climate change will now start going in reverse for a greener, serener Earth!

I'm so happy I could die right now

 

[dttprofessor]

Page 94 - Question - AMD Phoenix/Zen 4 APU Speculation and Discussion

Page 94 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Intel's world class engineering eked out a MASSIVE SIX EXTRA MINUTES of battery time from MTL H to U series.

Truly groundbreaking power efficiency!

Snapdragon Elite X stands no chance in challenging the reigning champion!

Humanity will rejoice! Global warming will be a thing of the past now! Climate change will now start going in reverse for a greener, serener Earth!

I'm so happy I could die right now

View attachment 97127

165H=165U,the same 7840H=7840U.

 

[uzzi38]

165H=165U,the same 7840H=7840U.

No they're not, 165H is 6+8 with 128EU iGPU, 165U is 2+8 with a 64EU iGPU

 

[poke01]

I do think X Elite will have better battery life even the 12 core part than any of Intel Arrow Lake mobile SoCs.

Lunar lake might beat the 12 core X Elite in terms of battery life but the lower core variants of X Elite like the 8 and 6 core will have better battery life than Lunar Lake.

 

[Gzxy]

I do think X Elite will have better battery life even the 12 core part than any of Intel Arrow Lake mobile SoCs.

Lunar lake might beat the 12 core X Elite in terms of battery life but the lower core variants of X Elite like the 8 and 6 core will have better battery life than Lunar Lake.

I believe ARL will beat X Elite in ST, MT. I don't know about efficiency in all TDP scenarios, it will depend on how good Lion and Skymont are. If you think about it Intel hasn't really changed it's P-cores arch since Golden Cove. It's more like small refreshes adding some cache and increasing clocks (except MTL which had clock regressions). I would say the same for Crestmont. They better deliver 35+% IPC on Lion Cove compared to Raptor cove. Also the ARL parts that will be based on MTL should be thrown out of in garbage cans. MTL is not competitive even right now where Zen 5 and X Elite are not out yet.

 

[igor_kavinski]

Also the ARL parts that will be based on MTL should be thrown out of in garbage cans.

Disagree. They should just sell them cheaper and kill the competition. That's what they are good at.

 

[SiliconFly]

Disagree. They should just sell them cheaper and kill the competition. That's what they are good at.

Yeah. MTL is actually good in efficiency (compared to prev gen). They should just reduce prices and hit the market with volume.

 

[igor_kavinski]

hit the market with volume.

That's the part they are having trouble with

 

[Ghostsonplanets]

They better deliver 35+% IPC on Lion Cove compared to Raptor cove.

... You're aware this is incredibly optimistic, right? Intel has been doing big changes in the past 5 years and that amounted for two jumps of 19%. And the early leaks around ARL show it barely being faster than RPL...

Also the ARL parts that will be based on MTL should be thrown out of in garbage cans.

MTL-U is still a good SoC for the average consumer and Intel can offer it at cheaper prices with good availability. The move to Intel 3 will also offer a 10% bump in energy efficiency.

Also, everything Arrow Lake will be based on Meteor Lake. It reuses the same tiles, with an updated tGPU tile to include XMX and the new Compute Tile.

 

[Ghostsonplanets]

https://twitter.com/x/status/1780556976839376904

None of these are news, but sharing anyway.

 

[igor_kavinski]

I'm really doubtful that the 4 LPE cores will have much, if any, contribution to LNL's MT score. If that's true, LNC looks to be a very potent P core.

 

[Ghostsonplanets]

From the slides Geddagod had saved from Yuuki_Ans, we can see that LNL had 2 internal SKU line-up. A Core 7 one, with full cache and Xe cores and a binned Core 5 one with less cache and a Xe core deactivated, with options for 16 or 32GB MOP memory.

Rest to know how Intel will market, consumer facing, the fanless and fanned SKUs.

 

[dullard]

Are we calling that "AI" now? Rathar than just looking something up?

The ability to understand my speech and respond back is AI. Especially when I just generically describe the item I want.

For example, the last time I went to Home Depot, I was told by a bonsai expert to get the 40% phosphorus plant food from Home Depot (and I had no idea of the brand). Well, there was no 40% phosphorus plant food in the garden center, plant area, and especially not in the plant food section. After tracking down and then speaking with 3 Home Depot employees, they had no clue and kept taking me to the plant food section. About an hour later, I found it in a hidden corner away far from any plant related items.

Or the last time I went to get a 8/2 AWG wire stripper. Not a single store employee in 4 stores had a clue that their 10/3 AWG wire strippers would not work. Harbor Freight's website listed it as in stock in the store. But it wasn't with any of the electrician supplies, or the wire stripper section, and I waited 30 minutes for two employees to search the back room. Eventually, I found it in an Icon brand hand tool grouping with pliers and wrenches.

It would have been far better for me to be able to say to a kiosk that I need 40% phosphorus plant food and have it just tell me where to go. Same with an 8/3 wire stripper. Since employees often have no clue of product details. Don't even get me started on asking them the difference between two brands of similar items when they don't even understand the basics of either.

 

[uzzi38]

Gaming is one use case. I think the biggest use cases would be industrial computers and kiosks (think any screen you interact with such as a help-desk, digital menu, information booth, point of sale systems, etc). There low power fanless applications are almost a must-have.

I'd really like them to use the AI features at stores at the end of aisles. I would love to be able ask where XYZ is located and have it say AISLE 18, lower shelf, bin #7.

I mean I think it'd be both lower effort to develop and much easier to run to just have an app with that information and a regular search bar on your mobile phone. And cheaper to the store as well, they don't have to fit a computer to the end of every aisle.

 

[SiliconFly]

View attachment 97147

@uzzi38, @mikk, @dullard, @tamz_msc, @HoveringStyle:

Tbh, I'm not a big fan of integrated GPU in general, so never bothered about these. But it looks like I'm the only one reading it wrong in this case!

Or am I?

This Golden Pig leak clearly says LNL has 8 Xe2 cores. And top MTL part has 8 Xe1 cores. (So, both are 128 EU I think).

And LNL (64 EU) roughly doubles the performance of MTL (64 EU) according to this SiSoftware Sandra benchmark.

Assuming it reflects well in the real world, aren't we looking at LNL top part (with 8 Xe2 cores) doubling the performance of MTL top part (with 8 Xe1 cores) ???

If true, then we can easily expect LNL top part (with 8 Xe2 cores) reach performance levels close to 3050 (at least 70% to 80% I'd say). Or have I overlooked something?

 

[tamz_msc]

@uzzi38, @mikk, @dullard, @tamz_msc, @HoveringStyle:

Tbh, I'm not a big fan of integrated GPU in general, so never bothered about these. But it looks like I'm the only one reading it wrong in this case!

Or am I?

This Golden Pig leak clearly says LNL has 8 Xe2 cores. And top MTL part has 8 Xe1 cores. (So, both are 128 EU I think).

And LNL (64 EU) roughly doubles the performance of MTL (64 EU) according to this SiSoftware Sandra benchmark.

Assuming it reflects well in the real world, aren't we looking at LNL top part (with 8 Xe2 cores) doubling the performance of MTL top part (with 8 Xe1 cores) ???

If true, then we can easily expect LNL top part (with 8 Xe2 cores) reach performance levels close to 3050 (at least 70% to 80% I'd say). Or have I overlooked something?

With integrated graphics it is seldom about the memory speeds or ALU count, and all about how much power can be shared with the iGPU in relation to the CPU cores.

 

[Ghostsonplanets]

@uzzi38, @mikk, @dullard, @tamz_msc, @HoveringStyle:

Tbh, I'm not a big fan of integrated GPU in general, so never bothered about these. But it looks like I'm the only one reading it wrong in this case!

Or am I?

This Golden Pig leak clearly says LNL has 8 Xe2 cores. And top MTL part has 8 Xe1 cores. (So, both are 128 EU I think).

And LNL (64 EU) roughly doubles the performance of MTL (64 EU) according to this SiSoftware Sandra benchmark.

Assuming it reflects well in the real world, aren't we looking at LNL top part (with 8 Xe2 cores) doubling the performance of MTL top part (with 8 Xe1 cores) ???

If true, then we can easily expect LNL top part (with 8 Xe2 cores) reach performance levels close to 3050 (at least 70% to 80% I'd say). Or have I overlooked something?

Nah, it's still 8 Xe Cores, the documents literally show that. But one Xe Core has only 8 EUs in Battlemage, but these EUs have double the Width. So Lunar Lake still has 8 Raytracing Units etc. because these are matched with Xe Cores.

As for performance, rumor mill says to expect similar or slightly better performance from LNL iGPU compared to MTL. But it needs to be remembered that LNL iGPU will be clocked lower and is also targeting a low power profile.

Granted, LNL performance at low power should be much, much better than MTL, which basically dies at 15W. But MTL will trump out over LNL due to being able to reach higher power targets.

LNL Battlemage PPC advancements, memory speed (8533x), MOP and the 8MB SLC might offset MTL power advantage, but it remains to be seen.

 

[uzzi38]

@uzzi38, @mikk, @dullard, @tamz_msc, @HoveringStyle:

Tbh, I'm not a big fan of integrated GPU in general, so never bothered about these. But it looks like I'm the only one reading it wrong in this case!

Or am I?

This Golden Pig leak clearly says LNL has 8 Xe2 cores. And top MTL part has 8 Xe1 cores. (So, both are 128 EU I think).

And LNL (64 EU) roughly doubles the performance of MTL (64 EU) according to this SiSoftware Sandra benchmark.

Assuming it reflects well in the real world, aren't we looking at LNL top part (with 8 Xe2 cores) doubling the performance of MTL top part (with 8 Xe1 cores) ???

If true, then we can easily expect LNL top part (with 8 Xe2 cores) reach performance levels close to 3050 (at least 70% to 80% I'd say). Or have I overlooked something?

I've literally seen the performance projections. Just like the slides from Igor on ARL, there's also one for LNL's iGPU compared to MTL-U (which caps out at 64EUs/4Xe cores). Even then Time Spy performance is less than a 2x gain (I'm not giving an exact number, I had too many people complain at me the last time I "leaked" something).

And again, those comparisons are against 15W MTL. Meteor Lake at 15W is in a totally different performance category to Phoenix/Hawk Point. It wouldn't even get close to 3050 levels of performance, regardless of if you're talking about desktop or mobile 3050.

 

[Philste]

Basically, BMG Xe Core = 2 Alchemist Xe Core, thus LNL 4 Xe² Core = 128 EUs.

Nah, it's still 8 Xe Cores, the documents literally show that. But one Xe Core has only 8 EUs in Battlemage, but these EUs have double the Width. So Lunar Lake still has 8 Raytracing Units etc. because these are matched with Xe Cores.