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	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.)

igor_kavinski · Dec 27, 2024

LENOVO 83KF vs HP HP OmniBook Ultra Laptop 14-fd0xxx - Geekbench

Intel shouldn't have any issues in the mobile space at least.

Battery life remains to be seen but for someone like me who is more interested in using laptops plugged in, that's pretty decent.

511 · Dec 27, 2024

igor_kavinski said:
LENOVO 83KF vs HP HP OmniBook Ultra Laptop 14-fd0xxx - Geekbench

Intel shouldn't have any issues in the mobile space at least.

Battery life remains to be seen but for someone like me who is more interested in using laptops plugged in, that's pretty decent.

I wished they made something more like LNL than ARL but i guess we will get that Q3 25 with Panther lake
Nvidia GPUs can kill battery life very quickly unless we get a new tech or we can simply disable the whole thing

hemedans · Dec 27, 2024

511 said:
I wished they made something more like LNL than ARL but i guess we will get that Q3 25 with Panther lake
Nvidia GPUs can kill battery life very quickly unless we get a new tech or we can simply disable the whole thing

Modern laptop do disable Dgpu and use igp most of time, it only use Dgpu when you need it like when you play games.

Markfw · Dec 27, 2024

igor_kavinski said:
LENOVO 83KF vs HP HP OmniBook Ultra Laptop 14-fd0xxx - Geekbench

Intel shouldn't have any issues in the mobile space at least.

Battery life remains to be seen but for someone like me who is more interested in using laptops plugged in, that's pretty decent.

Something other than geekbench 6 like a real review would be a better judge of the two.

511 · Dec 27, 2024

hemedans said:
Modern laptop do disable Dgpu and use igp most of time, it only use Dgpu when you need it like when you play games.

That's not true tbh i have one and the Nvidia GPU just turn on out of it's own you can compare any laptop with and without dGPU with same CPU and the dGPU one is worse in Battery Life

511 · Dec 27, 2024

Markfw said:
Something other than geekbench 6 like a real review would be a better judge of the two.

We will have to wait till CES perhaps

LightningZ71 · Dec 27, 2024

I'm still quite interested in finding out the performance of Arrow Lake-U/Meteor Lake refresh. I suspect it'll be an interesting comparison between that and other Intel mobile parts, especially if they wind up in laptops that will keep them in higher tdp levels...

RNR_Forte · Dec 27, 2024

511 said:
what would you gain by switching to ARM ?

Absolutely nothing

ARM still have the efficiency crown.

SiliconFly · Dec 27, 2024

RNR_Forte said:
ARM still have the efficiency crown.

Nope. This myth has been totally debunked. ARM ISA is NOT more efficient than x86 ISA. Even Jim Keller said the same. Some ARM processors tend to be more efficient than x86 processors. But that totally changed with Lunar Lake. Lunar Lake is more power efficient than its counterparts.

RNR_Forte · Dec 27, 2024

SiliconFly said:
Nope. This myth has been totally debunked. ARM ISA is NOT more efficient than x86 ISA. Some ARM processors tend to be more efficient than x86 processors. But that totally changed with Lunar Lake. Lunar Lake is more power efficient than its counterparts.

Yep. Zen family itself is pretty efficient too.

oak8292 · Dec 27, 2024

OneEng2 said:
Alternatives will indeed arrive (and already have). I don't disagree that this is where things are heading, only that there is quite some time before x86 will be in danger of being supplanted ..... at least in corporate machines.

There are just too many apps and IT dependencies .... and training, etc, etc that need to be changed over.

I think that in consumer laptops, things will change earlier ... but there will still likely be a pretty steep resistance to changing from x86. My own personal cost to change would be unbearable.

It will be a long time before x86 is supplanted but I think you are understating how far it already has been supplanted. Mac OS is about 15% of the Desktop/laptop user space out there.

Desktop Operating System Market Share Worldwide | Statcounter Global Stats

This graph shows the market share of desktop operating systems worldwide based on over 5 billion monthly page views.

gs.statcounter.com

After four years of Apple silicon that is probably 40-50% of the Mac OS switched over to ARM. One problem Mac OS had in corporate was central IT management software. Apple’s focus on consumers meant that they weren’t strong on developing IT tools. However Jamf is providing those tools in corporate environments and they have a lot of enterprise customers.

I would like to know where you got the 98% number for x86 in corporate. Even in enterprise I don’t believe it is that high. IBM Z systems have continued to maintain a presence. In cloud Amazon is on their fourth Graviton processor and may be using them for a lot of internal load.

The main advantage to ARM is the cost of IP if you can cover the NRE with adequate volume. AMD is still making money on x86 and they have about a 40% margin over fab margin. Anyone of the cloud providers with significant volume can potentially reduce costs by fabbing their own processor. ARM server design volume is still relatively small but V2 volumes are picking up which spreads design costs.

Intel is losing money on every processor as they can’t cover Capex costs. It will take a few years to see if moving to EUV actually reduces their costs adequately to be competitive. I don’t see them recovering architecture margin. So much of the high margin x86 product goes to cloud vendors who have volume for custom ARM processors and have switched their spend to accelerators (GPUs).

SiliconFly · Dec 27, 2024

RNR_Forte said:
Yep. Zen family itself is pretty efficient too.

Zen is not as efficient as Lunar Lake. Lunar Lake is pretty much equal to Apple M3 Silicon.

techjunkie123 · Dec 27, 2024

SiliconFly said:
Zen is not as efficient as Lunar Lake. Lunar Lake is pretty much equal to Apple M3 Silicon.

There's still a large gap between x86 and ARM when it comes to single core PPW. MT is similar, idle is similar with lunar lake, GPU is also similar. But ST loads still suck on x86, which sucks.

Now of course I'm not saying this is inherent to x86. Just something that hasn't been addressed by either AMD or Intel.

SiliconFly · Dec 27, 2024

techjunkie123 said:
There's still a large gap between x86 and ARM when it comes to single core performance. MT is similar, idle is similar with lunar lake, GPU is also similar. But ST loads still suck on x86, which sucks.

Now of course I'm not saying this is inherent to x86. Just something that hasn't been addressed by either AMD or Intel.

Nope. Lunar Lake ST is pretty much equivalent to M3.

GTracing · Dec 27, 2024

SiliconFly said:
Nope. Lunar Lake ST is pretty much equivalent to M3.

He meant single thread power consumption.

techjunkie123 · Dec 27, 2024

SiliconFly said:
Nope. Lunar Lake ST is pretty much equivalent to M3.

GTracing said:
He meant single thread power consumption.

Exactly, I meant ST power consumption. I edited my post.

ST PPW is still off by 400% or something....

Meteor Late · Dec 27, 2024

Lunar Lake is not as efficient as Apple M3. Keep in mind that Lunar Lake often times loses performance while on battery to achieve higher runtime battery numbers.
Testing laptop and CPU efficiency is difficult. But Apple has an order of magnitude higher efficiency than any x86 in terms of single core efficiency. Multicore efficiency is more about how many cores one crams into the chip, not exclusively of course, as there are different scaling strategies, but yeah.
Also, in terms of purely performance, Apple M3 still has higher single core than Lunar Lake.

SiliconFly · Dec 27, 2024

techjunkie123 said:
ST PPW is still off by 400% or something....

Nope. Thats just way off...

Meteor Late said:
But Apple has an order of magnitude higher efficiency than any x86 in terms of single core efficiency.

Nope. Lunar Lake isn't far off from M3.

Meteor Late said:
Also, in terms of purely performance, Apple M3 still has higher single core than Lunar Lake.

They're pretty much in the same class.

Meteor Late · Dec 27, 2024

SiliconFly said:
Nope. Thats just way off...

Nope. Lunar Lake isn't far off from M3.

They're pretty much in the same class.

Not really from what I've seen in Spec2017, in other benchmarks like Cinebench 2024 it's like 10% faster in SC compared to the much less common 268v, the difference is a bit higher against the 258v that is what you will find in almost any laptop with Lunar Lake out there, I tried to find one with 268v or 288v and I didn't find any.

techjunkie123 · Dec 27, 2024

SiliconFly said:
Nope. Thats just way off...

Nope. Lunar Lake isn't far off from M3.

They're pretty much in the same class.

Intel Lunar Lake CPU-Analyse - Der Core Ultra 7 258V enttäuscht bei Multi-Core-Leistung, punktet aber bei der Alltags-Effizienz

Notebookcheck CPU-Analyse des neuen Intel Lunar Lake Core Ultra 7 258V sowie Core Ultra 256V im Vergleich mit Snapdragon X Elite, AMD Zen 5 & Apple M3.

www.notebookcheck.com

M3 is about 15% faster than Lunar Lake while consuming less than half the power...Geekwans measurements are similar.

A 200-300% lead isn't trivial.

igor_kavinski · Dec 27, 2024

Apple silicon efficiency depends on the application's optimization. I'm using VLC player on M1 MBA to watch movies these days. Some file formats drain the battery a lot more while others let me watch three movies in a row. I'm not curious enough to use GSpot to figure out what codecs the offending files are using but it's a great example of how any hardware isn't magic if it's pushed outside of its intended way of operation.

ajsdkflsdjfio · Dec 27, 2024

techjunkie123 said:
Intel Lunar Lake CPU-Analyse - Der Core Ultra 7 258V enttäuscht bei Multi-Core-Leistung, punktet aber bei der Alltags-Effizienz

Notebookcheck CPU-Analyse des neuen Intel Lunar Lake Core Ultra 7 258V sowie Core Ultra 256V im Vergleich mit Snapdragon X Elite, AMD Zen 5 & Apple M3.

www.notebookcheck.com

M3 is about 15% faster than Lunar Lake while consuming less than half the power...Geekwans measurements are similar.

A 200-300% lead isn't trivial.

To get the entire picture you would need to compare the entire power curves of M3 vs Lion Cove, or at least compare results using the same power. You are comparing Lunar lake running at the high end of it's power curve with diminishing returns vs a single data point of the M3 running at lower power, likely at more of a sweet spot on it's power/perf curve. You're like comparing a 14900k running at 600W with + 15% perf vs a 14900k running at 300W, which, according to your logic, would make the 14900k have a 70% lead over itself.

Either way, if you compare the results at the same power M3 has around a 50% lead not 200% or 300%. Certainly not close, but it kind of makes sense given the wide design of M-series is tailor made for perf/watt and is more similar to ultra-efficient phone chips than desktop processors.

techjunkie123 · Dec 27, 2024

ajsdkflsdjfio said:
To get the entire picture you would need to compare the entire power curves of M3 vs Lion Cove, or at least compare results using the same power. You are comparing Lunar lake running at the high end of it's power curve with diminishing returns vs a single data point of the M3 running at lower power, likely at more of a sweet spot on it's power/perf curve. You're like comparing a 14900k running at 600W with + 15% perf vs a 14900k running at 300W, which, according to your logic, would make the 14900k have a 70% lead over itself.

Either way, if you compare the results at the same power M3 has around a 50% lead not 200% or 300%. Certainly not close, but it kind of makes sense given the wide design of M-series is tailor made for perf/watt and is more similar to ultra-efficient phone chips than desktop processors.

Nah.... Lunar Lake by default runs near the top of the power curve. Hence, this is the correct comparison point.

Otherwise you're looking at a core that's more than 50% slower at the same power limit. Might as well compare an efficiency core at that point....

The point is everyone is still waiting for the ultra efficient "phone" core from x86 that offers solid performance without compromising power efficiency too much.

ajsdkflsdjfio · Dec 27, 2024

techjunkie123 said:
Nah.... Lunar Lake by default runs near the top of the power curve. Hence, this is the correct comparison point.

Otherwise you're looking at a core that's more than 50% slower at the same power limit. Might as well compare an efficiency core at that point....

The point is everyone is still waiting for the ultra efficient "phone" core from x86 that offers solid performance without compromising power efficiency too much.

How are you so sure of that? How do you know Geekerwan didn't manually turn up the power of the core to get a wider range of data? In the first place Lunar-lake is a efficiency oriented chip, I doubt it'd be running at the most inefficient part of it's power curve 99% of the time, maybe in heavy 1t workloads for a short amount of time but otherwise probably not. Running at 12W per core in a 17W chip lol? Doesn't make sense.

BTW, in this discussion you are making the assumption that somehow M3 can't also ramp up it's power depending on the workload. You can literally see an M4 test Geekerwan ran which had a higher power limit close to lunar-lakes power limit, and in that scenario it's also ~50% faster not 2-3x.

Either way, without knowing the specific clock/power behavior of chips in various 1T/nT scenarios, it's probably best to compare power curve vs power curve, or at least compare results at the same power when you only have one data point.

techjunkie123 · Dec 27, 2024

ajsdkflsdjfio said:
How are you so sure of that? How do you know Geekerwan didn't manually turn up the power of the core to get a wider range of data? In the first place Lunar-lake is a efficiency oriented chip, I doubt it'd be running at the most inefficient part of it's power curve 99% of the time, maybe in heavy 1t workloads for a short amount of time but otherwise probably not. Running at 12W per core in a 17W chip lol? Doesn't make sense.

That's the whole point. 1T PPW sucks on x86. You can't clock lunar lake any higher lol it's already clocked to the moon. There's a reason why the top bin part isn't available.

ajsdkflsdjfio said:
BTW, in this discussion you are making the assumption that somehow M3 can't also ramp up it's power depending on the workload. You can literally see an M4 test Geekerwan ran which had a higher power limit close to lunar-lakes power limit, and in that scenario it's also ~50% faster not 2-3x.

Either way, without knowing the specific clock/power behavior of chips in various 1T/nT scenarios, it's probably best to compare power curve vs power curve, or at least compare results at the same power when you only have one data point.

MT is fine. That's the point. But 1T isn't. Time and time again.

There's a reason why x86 battery life sux under web browsing tests (while maintaining close to plugged in perf).

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

Senior member

Attachments

Lifer

Golden Member

Senior member

Moderator Emeritus, Elite Member

Golden Member

Golden Member

Platinum Member

Banned

Golden Member

Banned

Member

Golden Member

Member

Golden Member

Senior member

Member

Senior member

Golden Member

Senior member

Attachments

Member

Lifer

Member

Member

Member

Member