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.)

 

[cebri1]

Micro-Star International Co., Ltd. Claw A1M - Geekbench

Benchmark results for a Micro-Star International Co., Ltd. Claw A1M with an Intel Core Ultra 7 155H processor.

browser.geekbench.com

MSI Handheld with 155H. Really intrigued to see the performance and battery life on this one.

 

[TESKATLIPOKA]

Micro-Star International Co., Ltd. Claw A1M - Geekbench

Benchmark results for a Micro-Star International Co., Ltd. Claw A1M with an Intel Core Ultra 7 155H processor.

browser.geekbench.com

MSI Handheld with 155H. Really intrigued to see the performance and battery life on this one.

Let's welcome MSI CLAW.

It looks like It will have up to 32GB RAM, nice.

 

[uzzi38]

Micro-Star International Co., Ltd. Claw A1M - Geekbench

Benchmark results for a Micro-Star International Co., Ltd. Claw A1M with an Intel Core Ultra 7 155H processor.

browser.geekbench.com

MSI Handheld with 155H. Really intrigued to see the performance and battery life on this one.

Performance I can agree on, battery life with gaming handhelds is less about the chip involved and more about the power limits it runs at, which is usually 15W. So far MTL iGPU results at 15W have been a lot less impressive than that at 28W, but hopefully that can be fixed with future firmware improvements?

 

[DavidC1]

I am skeptical if the level of "IPC regression" is big as that test shows. This is one issue of only having mobile chips. You can never do a proper test. And even if you can have them at the same frequency, the desktop chip will outperform the mobile one by few %.

@Geddagod While you say both tests show similar regression, the E cores show a significant difference between the two tests.

Also, unless something is really messed up, scaling between chips do not vary that drastically, so much so that just 10% clock increase would result in IPC shifting upwards of 3-4%.

So far MTL iGPU results at 15W have been a lot less impressive than that at 28W, but hopefully that can be fixed with future firmware improvements?

Based on the curve it seems MTL/Intel 4 still has the same curve that existed since 10nm SF in the pursuit of ultra high frequencies. That means even with firmware improvements, at such low power levels it'll underperform.

The curve is a compromise between the two - if you want to reach the maximum frequency, Intel's way is the way to go. But you lose on the low power. To me, the rumors of Arrowlake having noticeably(even substantially) reduced clocks makes sense, because 1) it's unsustainable 2) it hurts mobile and server.

The curve can be optimized at both process and chip level. This is why the -U chips require more power at the same frequency compared to -H. The decision is deliberate. The -H needs extra clocks, while -U needs low leakage for battery life.

 

[eek2121]

Let's welcome MSI CLAW.

It looks like It will have up to 32GB RAM, nice.

I will stick with my Steam Deck.

These guys are forgetting the software part of the equation. The Steam Deck “just works” and it works well. Would I like to see Valve move to Zen 4 + RDNA3 or possibly Meteor Lake? Of course.

Actually those LPe cores in MTL would be great for running the SteamOS frontend/UI. Shoot, could run the entire OS on them…Even at 2.5ghz the *mont cores are performant.

 

[hemedans]

I will stick with my Steam Deck.

These guys are forgetting the software part of the equation. The Steam Deck “just works” and it works well. Would I like to see Valve move to Zen 4 + RDNA3 or possibly Meteor Lake? Of course.

Actually those LPe cores in MTL would be great for running the SteamOS frontend/UI. Shoot, could run the entire OS on them…Even at 2.5ghz the *mont cores are performant.

What's performance difference between E and LPe cores? Same IPC?

 

[H433x0n]

Based on the curve it seems MTL/Intel 4 still has the same curve that existed since 10nm SF in the pursuit of ultra high frequencies. That means even with firmware improvements, at such low power levels it'll underperform.

Crestmont seems to outperform everything at 5-6W (incl Zen 4). Disagree on the RWC curve too, its scaling (especially for 6VT) is a lot less linear.

I am worried about battery life for the MSI handheld though since it seems it falls off a cliff in gaming below 25W.

 

[mikk]

Performance I can agree on, battery life with gaming handhelds is less about the chip involved and more about the power limits it runs at, which is usually 15W. So far MTL iGPU results at 15W have been a lot less impressive than that at 28W, but hopefully that can be fixed with future firmware improvements?

Not just firmware, Golden pig said there is a test driver in the works with improved low power performance.

 

[DavidC1]

I don't know why you talk like Crestmont outperforming Zen 4 in perf/watt is a new thing. Alderlake's Gracemont did too, even with the heavy uncore. And I am not sure if anyone did MTL-like comparisons on mobile Alder/Raptor parts. Mobile chips would have lower overhead, thus a better comparison.

By the way that graph is representative of the pre-firmware update.

That's why we need a desktop, because laptops have all sort of unknowns that screws up the result. Even in desktops, it's only after dozens of reviews and months and months of arguing before people come to a consensus.

 

[DrMrLordX]

How are we still having a discussion about Meteor Lake vs. any competition? At least on a per-core basis, Meteor Lake should be the most advanced x86 mobile SoC on the market. It should clearly win ST performance versus its competition and older Intel SoCs, and it should use at least a little less power doing so. Without firmware updates and such. And yet here we are.

 

[cebri1]

How are we still having a discussion about Meteor Lake vs. any competition? At least on a per-core basis, Meteor Lake should be the most advanced x86 mobile SoC on the market. It should clearly win ST performance versus its competition and older Intel SoCs, and it should use at least a little less power doing so. Without firmware updates and such. And yet here we are.

It does use less power to achieve the same performance. Proven over and over again. Even in the first benchmarks. The pcore update just adds more margin between 13th gen and MTL.

 

[SiliconFly]

How are we still having a discussion about Meteor Lake vs. any competition? At least on a per-core basis, Meteor Lake should be the most advanced x86 mobile SoC on the market. It should clearly win ST performance versus its competition and older Intel SoCs, and it should use at least a little less power doing so. Without firmware updates and such. And yet here we are.

It definitely uses less power than previous gen to achieve similar performance.

 

[TESKATLIPOKA]

Crestmont seems to outperform everything at 5-6W (incl Zen 4).

I don't know why you talk like Crestmont outperforming Zen 4 in perf/watt is a new thing. Alderlake's Gracemont did too, even with the heavy uncore.

This is based on what?
Are you comparing a cluster of E-cores against a single core?

 

[H433x0n]

This is based on what?
Are you comparing a cluster of E-cores against a single core?

no, I was going off of David Huang's Specint data.

 

[jpiniero]

Hey, Dell finally has a Meteor Lake laptop available for sale. Only the i7 155H and starts at a grand. Delivery not for another 3 weeks though.

 

[TESKATLIPOKA]

no, I was going off of David Huang's Specint data.

You mean this?

Then Phoenix at 5W is comparable, at 6W is better. E-core at <5W is better.
I really want to know how he measured It.
I have PHX laptop and Ryzen Master is not working, the same for AMD Overdrive. Not sure about MTL and Extreme Tuning Utility.

 

[rainy]

I have PHX laptop and Ryzen Master is not working, the same for AMD Overdrive. Not sure about MTL and Extreme Tuning Utility.

Could you check, if this utility will work with your laptop?
https://amdaputuningutility.com/index.html

 

[DrMrLordX]

It does use less power to achieve the same performance. Proven over and over again. Even in the first benchmarks. The pcore update just adds more margin between 13th gen and MTL.

Except in the benchmarks where it uses more power. As much as I hate to say it, take a look at the Phoronix review. It's quite bad for Meteor Lake.

edit: also the perf/watt charts make Meteor Lake look pretty bad.

 

[TESKATLIPOKA]

Could you check, if this utility will work with your laptop?
https://amdaputuningutility.com/index.html

The app starts, that's already a good signal and It detects my APU.
Premade presets work, but you have to set Extreme Performance in Lenovo Vantage or you will be limited to max 30W power limit, regardless of which preset you choose.

Didn't play with It any more.
If MTL works I don't know, but for that you should use Intel's own APP considering this is mainly for AMD.

 

[tamz_msc]

As much as I hate to say it, take a look at the Phoronix review. It's quite bad for Meteor Lake.

One word: Linux.

Alder Lake also had sub-par performance at launch on Linux.

 

[cebri1]

Except in the benchmarks where it uses more power. As much as I hate to say it, take a look at the Phoronix review. It's quite bad for Meteor Lake.

edit: also the perf/watt charts make Meteor Lake look pretty bad.

No it didn't. Even in the first Notebookcheck review, which everyone used to tell us MTL was a failure, it beat the 13th gen at all tested power levels.

 

[eek2121]

It should clearly win ST performance versus its competition and older Intel SoCs, and it should use at least a little less power doing so.

Respectfully disagree.

If AMD/Intel released a chip tomorrow that had the same performance as Zen 4, but a half the power, I would call that a win.

If they managed to release a chip that was 10% slower, but consumed 1/3rd the power, that is still a win.

Cost, power consumption, and performance are all unique metrics that can independently make a successful product.

Meteor Lake isn’t perfect, but it is a win for Intel because it substantially improves perf/watt.

It is likely going to do well in terms of sales, especially if designs are compatible with Arrow Lake.

 

[coercitiv]

If AMD/Intel released a chip tomorrow that had the same performance as Zen 4, but a half the power, I would call that a win.

But Intel did not release a chip with the same performance as Zen 4 but half the power, did they?

 

[TESKATLIPOKA]

No it didn't. Even in the first Notebookcheck review, which everyone used to tell us MTL was a failure, it beat the 13th gen at all tested power levels.

View attachment 91525

He didn't compare It against Raptor Lake, or did he?
Was he even talking about Raptor vs MTL?

@cebri1: Yeah, I see It now.

 

[cebri1]

He didn't compare It against Raptor Lake, or did he?

"It should clearly win ST performance versus its competition and older Intel SoCs, and it should use at least a little less power doing so"