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

Tigerick · Aug 22, 2022

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.)

Hulk · Apr 28, 2024

coercitiv said:
We really should respect people's choices as long is does not affect the rest of us. In these communities it's a matter of etiquette, if not common sense.

I don't know if this is directed at me but I meant no disrespect. If I suddenly disappeared I would expect (hope) some of the people that I have been having in-depth discussions with for the past 20+ year might wonder what had happened to me. My post was one of respect and concern, not simple curiosity.

AMDK11 · Apr 28, 2024

I'm curious how Intel's "new" approach in LionCove will affect IPC's profit.

GoldenCove has 5x ALU, but it is 3x FP-ALU + 2x ALU. You can see from Zen4 how 4x ALU performs very well compared to GoldenCove, while having a separate scheduler and execution ports for the FPU.

LionCove has a separate schedule for 6x ALU and a separate schedule for 4 FPU ports.

Additionally, it has as many as 8 execution ports for memory operations, i.e. 6x AGU + 2x Store/Data (Zen5 4x AGU).

Hulk · Apr 28, 2024

AMDK11 said:
I'm curious how Intel's "new" approach in LionCove will affect IPC's profit.

GoldenCove has 5x ALU, but it is 3x FP-ALU + 2x ALU. You can see from Zen4 how 4x ALU performs very well compared to GoldenCove, while having a separate scheduler and execution ports for the FPU.

LionCove has a separate schedule for 6x ALU and a separate schedule for 4 FPU ports.

Additionally, it has as many as 8 execution ports for memory operations, i.e. 6x AGU + 2x Store/Data (Zen5 4x AGU).

Where did you get the Lion Cove architecture specifics? Is there a schematic out there somewhere?

AMDK11 · Apr 28, 2024

Hulk said:
Where did you get the Lion Cove architecture specifics? Is there a schematic out there somewhere?

Yes. The diagram is hidden in the LunarLake core image.

https://elchapuzasinformatico.com/wp-content/uploads/2024/04/Intel-Lunar-Lake-Dark-Silicon-Tile-Dummy.jpg

LionCove P core:
8-Way Dispatch/Rename

a LionCove diagram can be compared to a RedwoodCove or GoldenCove diagram.

There seems to be a 4-Way predictor.

In the place of decoding and sending from the UOP cache, I counted 24-26(?) items. GoldenCove has 14, i.e. 6 from the decoder and 8 from the uop cache.

in total, LionCove has 18 execution ports compared to 12 of GoldenCove.

under the AGU+S/D units there are three blocks suggesting the division of L2 into 512KB + 2.5MB(?).

The L3 cache appears to be of a different design.

The Skymont core diagram shows 3x 3-Way decoding(Gracemont-Crestmont 2x 3-Way).

I'm sure Intel will present the same diagrams and descriptions at the LionCove and Skymont presentations.

SiliconFly · Apr 28, 2024

Hulk said:
My post was one of respect and concern...

I thought so too. Wonder why it was misinterpreted!

coercitiv · Apr 28, 2024

Hulk said:
I don't know if this is directed at me but I meant no disrespect.

Definitely not directed at you, sorry if it may have seemed that way.

S'renne · Apr 28, 2024

AMDK11 said:
Yes. The diagram is hidden in the LunarLake core image.

https://elchapuzasinformatico.com/wp-content/uploads/2024/04/Intel-Lunar-Lake-Dark-Silicon-Tile-Dummy.jpg

LionCove P core:
8-Way Dispatch/Rename

a LionCove diagram can be compared to a RedwoodCove or GoldenCove diagram.

There seems to be a 4-Way predictor.

In the place of decoding and sending from the UOP cache, I counted 24-26(?) items. GoldenCove has 14, i.e. 6 from the decoder and 8 from the uop cache.

in total, LionCove has 18 execution ports compared to 12 of GoldenCove.

under the AGU+S/D units there are three blocks suggesting the division of L2 into 512KB + 2.5MB(?).

The L3 cache appears to be of a different design.

The Skymont core diagram shows 3x 3-Way decoding.

I'm sure Intel will present the same diagrams and descriptions at the LionCove and Skymont presentations.

That's actually pretty interesting

AMDK11 · Apr 28, 2024

Here is the original in higher resolution:

https://cdn.wccftech.com/wp-content/uploads/2024/04/Intel-Lunar-Lake-Core-Ultra-200-CPUs-_2.png

Hulk · Apr 28, 2024

AMDK11 said:
Here is the original in higher resolution:

https://cdn.wccftech.com/wp-content/uploads/2024/04/Intel-Lunar-Lake-Core-Ultra-200-CPUs-_2.png

Dead link. Anyone have the high resolution photo? I really want to see this.

mikk · Apr 28, 2024

Hulk said:
Dead link. Anyone have the high resolution photo? I really want to see this.

Should be from here: https://wccftech.com/intel-lunar-lake-core-ultra-200-cpus-over-100-ai-tops-3x-npu-performance/
Or here: https://www.hardwareluxx.de/index.p...ntel-nennt-100-tops-an-ai-rechenleistung.html

It's not really high resolution though.

AMDK11 · Apr 28, 2024

I quickly described the LionCove diagram, but it does not contain all the data, and the photo is not of high resolution, so you need to make sure by looking at the neighboring cores.

SiliconFly · Apr 28, 2024

AMDK11 said:
Yes. The diagram is hidden in the LunarLake core image.

LionCove P core:
8-Way Dispatch/Rename

a LionCove diagram can be compared to a RedwoodCove or GoldenCove diagram.

There seems to be a 4-Way predictor.

In the place of decoding and sending from the UOP cache, I counted 24-26(?) items. GoldenCove has 14, i.e. 6 from the decoder and 8 from the uop cache.

in total, LionCove has 18 execution ports compared to 12 of GoldenCove.

under the AGU+S/D units there are three blocks suggesting the division of L2 into 512KB + 2.5MB(?).

The L3 cache appears to be of a different design.

The Skymont core diagram shows 3x 3-Way decoding(Gracemont-Crestmont 2x 3-Way).

I'm sure Intel will present the same diagrams and descriptions at the LionCove and Skymont presentations.

Excellent 👍

Saylick · Apr 28, 2024

Inb4 WTFTech writes a new article on this.

If any of their authors is in these threads and sees this, you can go suck it!

AMDK11 said:
Yes. The diagram is hidden in the LunarLake core image.

https://elchapuzasinformatico.com/wp-content/uploads/2024/04/Intel-Lunar-Lake-Dark-Silicon-Tile-Dummy.jpg

LionCove P core:
8-Way Dispatch/Rename

a LionCove diagram can be compared to a RedwoodCove or GoldenCove diagram.

There seems to be a 4-Way predictor.

In the place of decoding and sending from the UOP cache, I counted 24-26(?) items. GoldenCove has 14, i.e. 6 from the decoder and 8 from the uop cache.

in total, LionCove has 18 execution ports compared to 12 of GoldenCove.

under the AGU+S/D units there are three blocks suggesting the division of L2 into 512KB + 2.5MB(?).

The L3 cache appears to be of a different design.

The Skymont core diagram shows 3x 3-Way decoding(Gracemont-Crestmont 2x 3-Way).

I'm sure Intel will present the same diagrams and descriptions at the LionCove and Skymont presentations.

Looks like 8-way decode for LNC? On the Mop cache side, it looks like an additional 12?

AMDK11 · Apr 28, 2024

AMDK11 · Apr 28, 2024

Saylick said:
Inb4 WTFTech writes a new article on this.

If any of their authors is in these threads and sees this, you can go suck it!

Looks like 8-way decode for LNC? On the Mop cache side, it looks like an additional 12?
View attachment 98053

View attachment 98054

Note that Intel describes this part as a predictor in the diagram.

And the MSROM may have been moved higher next to L1-I to accommodate all the decoding and sending entries from the uop cache.

Henry swagger · Apr 28, 2024

AMDK11 said:
Note that Intel describes this part as a predictor in the diagram.

And the MSROM may have been moved higher next to L1-I to accommodate all the decoding and sending entries from the uop cache.

Lion cove is 8 wide like apple's firestorm core.. hope they get high ipc at low clocks too

S'renne · Apr 28, 2024

Henry swagger said:
Lion cove is 8 wide like apple's firestorm core.. hope they get high ipc at low clocks too

Seems like the intention for Lion Cove onwards is following a similar path as Apple over AMD in core designs. Wasn't it Kepler or whoever it was suggesting that AMD has higher power draw and drastic performance improvement whereas Intel is finally going for sane power usage, maybe even lower with Lion Cove with performance uplift despite lower clocks?

Joe NYC · Apr 28, 2024

S'renne said:
Seems like the intention for Lion Cove onwards is following a similar path as Apple over AMD in core designs. Wasn't it Kepler or whoever it was suggesting that AMD has higher power draw and drastic performance improvement whereas Intel is finally going for sane power usage, maybe even lower with Lion Cove with performance uplift despite lower clocks?

I think this is relative to the starting point (previous gen) not relative to competition.

Henry swagger · Apr 29, 2024

S'renne said:
Seems like the intention for Lion Cove onwards is following a similar path as Apple over AMD in core designs. Wasn't it Kepler or whoever it was suggesting that AMD has higher power draw and drastic performance improvement whereas Intel is finally going for sane power usage, maybe even lower with Lion Cove with performance uplift despite lower clocks?

Yeah.. apple really shows high clocks are useless if you have a good architecture.. apple is the gold standard in efficiency on laptop

H433x0n · Apr 30, 2024

This video will surely start an entertaining conversation.

Place your bets as to how fellow forum contributors will respond.. the options are listed below:

A) paid for shill review
B) Measurement error
C) Unfair comparison
D) All of the above

Markfw · Apr 30, 2024

H433x0n said:
This video will surely start an entertaining conversation.

Place your bets as to how fellow forum contributors will respond.. the options are listed below:

A) paid for shill review
B) Measurement error
C) Unfair comparison
D) All of the above

None of the above.... IMO, they both had places where they won, but overall it really was pretty even. Thats the part I disagree with the reviewer on. But I will agree that its a stopgap by both for next years Zen 5 vs Lunar lake that we should wait for.

gdansk · Apr 30, 2024

H433x0n said:
This video will surely start an entertaining conversation.

Place your bets as to how fellow forum contributors will respond.. the options are listed below:

A) paid for shill review
B) Measurement error
C) Unfair comparison
D) All of the above

Here you go: in the real world it looks easier to configure Phoenix refresh than MTL (which so many vendors have messed up now).

Hitman928 · Apr 30, 2024

H433x0n said:
This video will surely start an entertaining conversation.

Place your bets as to how fellow forum contributors will respond.. the options are listed below:

A) paid for shill review
B) Measurement error
C) Unfair comparison
D) All of the above

Only thing I'll say about that review is their battery life test for "internet browsing" is ridiculous. I've commented on it before but it's probably the worst browsing test out there. Anyway, here are some additional reviews comparing both AMD and Intel models that show things in a bit different light. Make of it as you will.

2024 Asus Zenbook 14 OLED (UM3406HA): DOES AMD BURY INTEL?

Asus Zenbook 14 OLED (UM3406HA): https://goto.walmart.com/c/1441847/1398372/16662?sourceid=imp_000011112222333344&veh=aff&u=https%3A%2F%2Fwww.walmart.com%2Fi...

www.youtube.com

Abwx · Apr 30, 2024

Hitman928 said:
Only thing I'll say about that review is their battery life test for "internet browsing" is ridiculous. I've commented on it before but it's probably the worst browsing test out there. Anyway, here are some additional reviews comparing both AMD and Intel models that show things in a bit different light. Make of it as you will.

2024 Asus Zenbook 14 OLED (UM3406HA): DOES AMD BURY INTEL?

Asus Zenbook 14 OLED (UM3406HA): https://goto.walmart.com/c/1441847/1398372/16662?sourceid=imp_000011112222333344&veh=aff&u=https%3A%2F%2Fwww.walmart.com%2Fi...

www.youtube.com

In the Hardware Canucks fake review the AMD APU allegedly at 41W has 1h26 battery life at full load, while the MTL at allegedly 34W has 1h17, and that s with a same battery.
Obviously the MTL laptop use more power in the benches, Hardware Canucks are tricking the numbers.

Same for the higher temperature of PHX, they do no state the fan noise but it makes no doubt that it spin much harder in the MTL laptop, but since PHX has a higher temp they make people believe that it consume more, that s how you deceive the masses.

Geddagod · Apr 30, 2024

Hitman928 said:
Only thing I'll say about that review is their battery life test for "internet browsing" is ridiculous. I've commented on it before but it's probably the worst browsing test out there. Anyway, here are some additional reviews comparing both AMD and Intel models that show things in a bit different light. Make of it as you will.

2024 Asus Zenbook 14 OLED (UM3406HA): DOES AMD BURY INTEL?

Asus Zenbook 14 OLED (UM3406HA): https://goto.walmart.com/c/1441847/1398372/16662?sourceid=imp_000011112222333344&veh=aff&u=https%3A%2F%2Fwww.walmart.com%2Fi...

www.youtube.com

Similar battery life between MTL and PHX/HWK for these laptop models it looks like.

Abwx said:
In the Hardware Canucks fake review the AMD APU allegedly at 41W has 1h26 battery life at full load, while the MTL at allegedly 34W has 1h17, and that s with a same battery.
Obviously the MTL laptop use more power in the benches, Hardware Canucks are tricking the numbers.

Same for the higher temperature of PHX, they do no state the fan noise but it makes no doubt that it spin much harder in the MTL laptop, but since PHX has a higher temp they make people believe that it consume more, that s how you deceive the masses.

Jeezus the cope is sad. Not everything is a conspiracy.

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

Senior member

Attachments

Diamond Member

Senior member

Diamond Member

Senior member

Golden Member

Diamond Member

Member

Senior member

Diamond Member

Diamond Member

Senior member

Golden Member

Diamond Member

Senior member

Senior member

Senior member

Member

Platinum Member

Senior member

Senior member

Moderator Emeritus, Elite Member

Platinum Member

Diamond Member

Lifer

Golden Member