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.)

 

[SiliconFly]

.

 

[Khato]

Including the DRAM on package is a design choice, it also brings important advantages not just another contributor to package power. You get slightly better memory tuning, you get considerably lower mem controller voltage and Intel also gets to ensure all chips have LPDDR5X 8533. The price they pay by lowering max available package power is arguably more than made up by higher memory performance in combination with around 100mV less voltage needed to drive the memory chips.

Nobody brings up the advantage of having mem on package when comparing battery life, because it would be absurd to penalize LNL for an advantageous design choice. The same applies when scaling power though, it would be absurd to penalize Strix, MTL or even the upcoming ARL-U for their design choices.

Not sure what the point of the above is?

My point is that if you're trying to compare CPU + iGPU efficiency by providing same power to different designs, then the same power budget should be provided to the CPU + iGPU. LNL has about 1.5-2W less power available to the CPU + iGPU if set to the same PL numbers due to the inclusion of DRAM in those numbers, no?

 

[Hitman928]

Are we unironically arguing that battery life doesn’t matter? If we’re using this argument, isn’t RPL a superior mobile product over Phoenix & Hawk Point? It had more cores, better multi threading and a slight edge in single threading.

There definitely comes a point where a person doesn’t care about additional battery life, but there are plenty of people out there for whom that line is still far away and they will take all the battery life they can get.

 

[OneEng2]

I've been hearing this since the Llano days. "Wait for Trinity, that's the real competitive laptop part!"

AMD just don't have the connections and the volume, and I don't know why at this point. Their APUs have been excellent for years now, but they're still losing out to Intel left right and center. My only conclusion is that Intel can out manufacture them with the sheer capacity of their fabs, while AMD has to fight Nvidia and Apple for space.

Yet Intel's bottom line is bleeding red while AMD's books are flush. Can you explain this?

18A looks to be a lot better than TSMC N3 and very close to N2. Also, 18A is expected to have higher performance than N2 because of BSPDN (higher Fmax than N2, cos N2 doesn't have BSPDN).

Not what I have read anywhere. 18A looks to be about on par with N3P.... which isn't a bad thing for Intel who have been horridly behind. It is not expected to be competitive with N2 and I believe even Intel has stated this IIRC.

View attachment 108311

No more free hall pass for Turin. Looks like it’s gonna be a tough fight after all.

Looks like .

My feeling is that with full Zen 5 cores and better power efficiency, Turin 128 will likely exceed Granite Rapids by >30% ... but no longer a bloodbath.

 

[coercitiv]

Not sure what the point of the above is?

My point is that if you're trying to compare CPU + iGPU efficiency by providing same power to different designs, then the same power budget should be provided to the CPU + iGPU. LNL has about 1.5-2W less power available to the CPU + iGPU if set to the same PL numbers due to the inclusion of DRAM in those numbers, no?

LNL has a different design, the mem on package is a tradeoff that introduces both a penalty and a series of advantages. One cannot ask for the penalty to be nullified while the advantages are still in effect.

Imagine asking for LNL to run higher uncore voltages during battery life tests because MTL does not have mem on package. Sounds absurd enough to get the point across?

 

[AcrosTinus]

Yet Intel's bottom line is bleeding red while AMD's books are flush. Can you explain this?

Not what I have read anywhere. 18A looks to be about on par with N3P.... which isn't a bad thing for Intel who have been horridly behind. It is not expected to be competitive with N2 and I believe even Intel has stated this IIRC.

Looks like .

My feeling is that with full Zen 5 cores and better power efficiency, Turin 128 will likely exceed Granite Rapids by >30% ... but no longer a bloodbath.

Exceeding Granite Rapids in what ? By 30% or more, people consider such lofty predictions a "bloodbath".

The Zen5 cores are great but they are not 1.5 Generations better. The IPC is similarish, maybe due to the core being more efficient they might clock higher in the same 500W envelope.

 

[SiliconFly]

.

 

[yuri69]

AMD basically goes one server gen per 2 years. Intel goes one per year - 2023 Sapphire + Emerald; 2024 Granite; 2025 Diamond

Intel is gonna be server leader from 2025 Diamond Rapids launch 'till late 2026 Venice launch.

 

[cebri1]

Well SPR, EMR and GNR they are refinements of each other for the most part.

 

[yuri69]

Well SPR, EMR and GNR they are refinements of each other for the most part.

Seeing the performance gains, I would say these refinements are pretty good.

 

[adroc_thurston]

Ryzen AI 7/9 says Hi.

?

Yes?

No. Smaller chassis means cheaper.

I think you might be underestimating how AI obsessed the industry is right now.

That's silly marketing. Not related to the actual market segmentation stuff.

but they're still losing out to Intel left right and center

AMD is at record ~23% laptop MSS. Come again?

 

[cebri1]

Seeing the performance gains, I would say these refinements are pretty good.

Well SPR was horrible lol. EMR was a very nice jump considering it was built on the same core and the same node, but with a fixed memory subsystem.

GNR is more of a jump, due to RWC+ (Still GLC++++) & Intel 3.

DMR should be what's propels Intel ahead. 18A plus first really new CPU architecture since Golden Cove on server.

 

[Josh128]

So its been established that display has a huge bearing on battery life then, no?

Heres how the same Vivobook S 14 with HX 370 faired in runtimes vs the super-duper ARM- beating efficiency Core Ultra 2 258V. Whats that, Strix gets
10% better average runtime despite having only 4% larger capacity battery? How can that be?? Maybe display type and size does matter after all.

 

[CouncilorIrissa]

Lion Cove: Intel’s P-Core Roars

Intel’s mobile CPUs have undergone massive changes over the past couple generations as Intel defends its laptop market against AMD, Qualcomm, and to a lesser extent Apple.

chipsandcheese.com

 

[cebri1]

So its been established that display has a huge bearing on battery life then, no?

Heres how the same Vivobook S 14 with HX 370 faired in runtimes vs the super-duper ARM- beating efficiency Core Ultra 2 258V. Whats that, Strix gets
10% better average runtime despite having only 4% larger capacity battery? How can that be?? Maybe display type and size does matter after all.

View attachment 108348

Yes it has been established by every single reviewer. Stop polluting the thread. And no, display size from 14 to 16 barely affects power consumption.

 

[poke01]

That Jarrod video is not good at all. Many shortcomings and his results are different to many other reviewers because of this.

 

[Hitman928]

That Jarrod video is not good at all. Many shortcomings and his results are different to many other reviewers because of this.

What are the shortcomings?

 

[adroc_thurston]

display size from 14 to 16 barely affects power consumption.

?
Yeah it does, especially for OLEDs.
There's a good reason OLED TVs aren't efficient at all.

 

[cebri1]

?
Yeah it does, especially for OLEDs.
There's a good reason OLED TVs aren't efficient at all.

2 more inches, 5W more? 7-8W tops. And the 16 inch laptop has a larger batery as well.

However it's nice to see that the battery debate, that yesterday was stupid because there are power outlets everywhere, today is relevant again.

But let's keep polluting the thread.

edit: btw, 5-8W on average. If you test an OLED laptop, depending on the content, it can even be more efficient that an IPS.

 

[SiliconFly]

.

 

[SiliconFly]

.

 

[adroc_thurston]

5W more? 7-8W tops

That's a ton of power, big dawg.

If you test an OLED laptop, depending on the content, it can even be more efficient that an IPS.

It's only ever more efficient at absolutely miserable APL levels which is not laptop usage model at all!
Hope that helps.

 

[cebri1]

That's a ton of power, big dawg.

It's only ever more efficient at absolutely miserable APL levels which is not laptop usage model at all!
Hope that helps.

The issue is that we are seeing differences of 20W under load and people are attributing it to the OLED display.

 

[poke01]

What are the shortcomings?

Use of different power profiles and using XeSS for Wukong instead of FSR

 

[adroc_thurston]

The issue is that we are seeing differences of 20W under load and people are attributing it to the OLED display.

That's just PPT difference and is an OEM preference.