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.)

 

[Elfear]

If these are the final benchmark results (or close to), I just don't know what the point of LNL is besides making for a good handheld gaming PC chip (but likely an expensive one).

You want something that's good at low power? M series chips destroy it even fanless, and first gen Snapdragon X Elite handily beats it when thermally limited to 20W. Even power limiting Strix to 15W or around there is likely to run circles around it.

And this isn't gonna be a cheap chip, the packaging is advanced, it's N3B, ... So who is this for? Is the entire pitch the fact that it targets 17W on x86 and isn't completely terrible?

For the crowd who don't want to play in Apple's garden (raises hand) and also want something light, power-efficient, and relatively powerful. We'll have to see what AMD brings to the table to compete but Lunar looks to be pretty decent.

 

[dullard]

And this isn't gonna be a cheap chip, the packaging is advanced, it's N3B, ... So who is this for? Is the entire pitch the fact that it targets 17W on x86 and isn't completely terrible?

Many of us will never use Apple products (for me, I'd have to strongly weigh the options if my life depended on it). They don't even run my necessary software without crashes and slow emulation--but, I'm not a typical user either.

It'll have great GPU, great AI capability, low power, long enough battery life, and as you said it will have okay CPU performance (you just don't get high CPU performance on 9 to 30 W from ANY company). Perfect for kiosks (think AI powered help desk, information booths, signs automatically pointing to your favorite food/gate as you run through the airport), true inspection on factory lines, entertainment, daily use around the house/on-the-go, etc. Something that you just toss in your purse or throw in your backpack. If you want a workhorse number-crunching powerhouse, then this isn't it.

 

[Hitman928]

But based on this, it doesn't really do that. It barely competes with M2, doesn't get anywhere close to M3, and it will launch against M4. I'm not even saying it has to be as good as M4, but be less than 2 gens behind when it launches should have been an attainable goal.

So if the goal was to send the message that they can't compete with Apple Silicon at all but can make a product that's sorta OK in the segment, then those scores are fine but if not they're worrying.

If LNL can offer a significant step up in battery life over current x86 offerings, it will have a lot of value for thin and light notebooks and maybe tablets.

Many of us will never use Apple products (for me, I'd have to strongly weigh the options if my life depended on it). They don't even run my necessary software without crashes and slow emulation--but, I'm not a typical user either.

It'll have great GPU, great AI capability, low power, long enough battery life, and as you said it will have okay CPU performance (you just don't get high CPU performance on 9 to 30 W from ANY company). Perfect for kiosks (think AI powered help desk), entertainment, daily use around the house/on-the-go, etc. Something that you just toss in your purse or throw in your backpack. If you want a workhorse number-crunching powerhouse, then this isn't it.

Almost all kiosks use dirt cheap processors, LNL will be too expensive for them.

 

[FlameTail]

So will 16 GB retail LNL chips also have single-rank memory?

 

[AcrosTinus]

But based on this, it doesn't really do that. It barely competes with M2, doesn't get anywhere close to M3, and it will launch against M4. I'm not even saying it has to be as good as M4, but be less than 2 gens behind when it launches should have been an attainable goal.

So if the goal was to send the message that they can't compete with Apple Silicon at all but can make a product that's sorta OK in the segment, then those scores are fine but if not they're worrying.

For 8C/8T with hopefully good battery life, it is very nice an upgrade to the best (imho.Tiger Lake G7). x86 is and will always be about compatibility the "just works" for fleets of notebooks. You will not have issues with Davinci Resolve, the Adobe package, you know what you get.
Outside of benchmarks the MT performance being max. 40% behind will never be noticed. Intel's focus here with the 200V series is battery life and mainstream performance. For more you might have to look at the Arrow Lake Notebook series which will have Skymont on the ring.

 

[uzzi38]

We don’t know how Strix performs at <20W, I’m also not sure if SDXE would outperform it under 20W either.

Why not wait until we get a bit more info before declaring anything dead? It’s a different segment than both Strix and SDXE. This is designed to compete with M2 or M3 in MacBook Air with both of them being 8C/8T designs. For chips that have more multicore performance that scales with power there’s ARL-U (MTL on Intel 3) and ARL-H.

If you're expecting Strix to perform worse than Phoenix below 20W you're probably setting yourself up to fail tbh. Like I said earlier - at least in R23 - Phoenix sits in a similar performance bracket to LNL at 15W and seemingly has an increasingly large margin the higher you go afterwards.

That being said:

If these are the final benchmark results (or close to), I just don't know what the point of LNL is besides making for a good handheld gaming PC chip (but likely an expensive one).

You want something that's good at low power? M series chips destroy it even fanless, and first gen Snapdragon X Elite handily beats it when thermally limited to 20W. Even power limiting Strix to 15W or around there is likely to run circles around it.

And this isn't gonna be a cheap chip, the packaging is advanced, it's N3B, ... So who is this for? Is the entire pitch the fact that it targets 17W on x86 and isn't completely terrible?

Given the idle power figures shared by the poster of those benchmark runs, I expect LNL to come close to or probably even exceed the battery life of SDXE. So if you are interested in that level of battery life while sticking to an x86 platform (and trust me, when you see runs of GB5 through Prism you'll really consider sticking to x86 as a real positive) then LNL is the chip to get.

Seriously, numbers like these are utterly insane. This is MTL-U vs LNL package power... Which includes memory for LNL but not for MTL. I would genuinely be extremely surprised if LNL loses to SDXE on battery life, to be frank.

That's a huge selling point for thin and light devices.

 

[Hitman928]

If you're expecting Strix to perform worse than Phoenix below 20W you're probably setting yourself up to fail tbh. Like I said earlier - at least in R23 - Phoenix sits in a similar performance bracket to LNL at 15W and seemingly has an increasingly large margin the higher you go afterwards.

That being said:



Given the idle power figures shared by the poster of those benchmark runs, I expect LNL to come close to or probably even exceed the battery life of SDXE. So if you are interested in that level of battery life while sticking to an x86 platform (and trust me, when you see runs of GB5 through Prism you'll really consider sticking to x86 as a real positive) then LNL is the chip to get.

Seriously, numbers like these are utterly insane. This is MTL-U vs LNL package power... Which includes memory for LNL but not for MTL. I would genuinely be extremely surprised if LNL loses to SDXE on battery life, to be frank.

That's a huge selling point for thin and light devices.

Not sure I believe his numbers, but he said MTL numbers include memory as well.

 

[uzzi38]

Not sure I believe his numbers, but he said MTL numbers include memory as well

Lets just say there hypothetically may have been an old something or another that looks very similar, and in the top right corner it said something "SoC +16GB memory", but over to the left it said "SoC" multiple times.

Hypothetically speaking. Of course something like this couldn't possibly be real!

 

[H433x0n]

But based on this, it doesn't really do that. It barely competes with M2, doesn't get anywhere close to M3, and it will launch against M4. I'm not even saying it has to be as good as M4, but be less than 2 gens behind when it launches should have been an attainable goal.

So if the goal was to send the message that they can't compete with Apple Silicon at all but can make a product that's sorta OK in the segment, then those scores are fine but if not they're worrying.

The LNL ES scores slot it in between M2 & M3 in both 1T and nT.

Nobody is able to compete with Apple at the moment. Your criticism could also be leveled at AMD who didn’t even attempt to go after this segment and Qualcomm whom had a good first showing but ultimately fell short as well.

I would argue the closest you can get to the M2/M3 within the Microsoft landscape would be LNL. Remains to be seen how big this market actually is, guess we’ll find out soon enough.

 

[The Hardcard]

Wintel for the masses is still the default. Much less so for tech bros, but for people uninterested in computers, but need one to get things done, Intel is the only one that doesn’t need to wow. All the others need to wow. Maybe less so for AMD.

Outside a certain fanbase, Apple needs to wow. They have a certain cachet, but it comes at a price. The M1 pulled a wow they got a few extra people who don’t otherwise care to buy MacBooks. They can probably still ride that wave, but the gap is not so big now. 12 hours unplugged versus 5 hours is huge. 18 hours versus 14 hours is not enough to flip many to a laptop they don’t otherwise want.

Lunar Lake can be successful by just being in the pack, and so far it looks like it will be. It doesn’t need exceptional performance or exceptional battery life It just needs to not get its ass kicked on any metric. That alone will give it first place in sales and market share for thin and lights.

Frankly, this is really it can’t lose situation for a laptop buyers. You should really just be able to get the laptop you like with the OS and software you want to use. You will get similar single threaded and multi threaded performance and battery life will be in the same ballpark regardless of the chip inside.

 

[dullard]

Almost all kiosks use dirt cheap processors, LNL will be too expensive for them.

And how many dirt cheap processors can actually do a respectable job of replacing human labor for advise at desk? A kiosk with AI tuned for the specific application is nothing cost wise in comparison.

 

[dullard]

So will 16 GB retail LNL chips also have single-rank memory?

That is what the rumors state so far.

 

[poke01]

you just don't get high CPU performance on 9 to 30 W from ANY company

I mean you do with M4.

They don't even run my necessary software without crashes and slow emulation--but, I'm not a typical user either.

what software do you use?

 

[poke01]

Seriously, numbers like these are utterly insane. This is MTL-U vs LNL package power... Which includes memory for LNL but not for MTL. I would genuinely be extremely surprised if LNL loses to SDXE on battery life, to be frank.

Yep they really are, compared with Meteor Lake. Lunar is so good. I would say that Lunar will also beat Strix in battery life.

there could still be more improvements for Lunar as still in ES stage. There is so much to compare packaging, PDN, core designs. This is the most interesting thing for me from Intel. It’s packed with new IP.

Which includes memory for LNL but not for MTL

It includes memory for MTL as well.

https://twitter.com/x/status/1806157134805062036

 

[TwistedAndy]

If these are the final benchmark results (or close to), I just don't know what the point of LNL is besides making for a good handheld gaming PC chip (but likely an expensive one).

You want something that's good at low power? M series chips destroy it even fanless, and first gen Snapdragon X Elite handily beats it when thermally limited to 20W. Even power limiting Strix to 15W or around there is likely to run circles around it.

And this isn't gonna be a cheap chip, the packaging is advanced, it's N3B, ... So who is this for? Is the entire pitch the fact that it targets 17W on x86 and isn't completely terrible?

Lunar Lake is expected to be more efficient than Apple M3 and nearly the same in terms of performance. It's expected to be an excellent platform for thin and light devices.

 

[poke01]

Lunar Lake is expected to be more efficient than Apple M3

Can we not make these false statements it’s already has worse performance per watt from the ES benchmarks than M3. Lunar needed to clock at 4.8GHz to get a score of 2800 whereas M3 does 3200 at 4.06GHz.

The M3 P core doesn’t clock as high so more efficient there and also M3 has LITTLE’s. I really don’t know why you would think Lunar would be more efficient.

Compared to Meteor lake it will be efficient.

 

[TwistedAndy]

Can we not make these false statements it’s already has worse performance per watt from the ES benchmarks than M3. Lunar needed to clock at 4.8GHz to get a score of 2800 whereas M3 does 3200 at 4.06GHz.

Apple M3 scores nearly 3000-3100 in Geekbench, M4 - nearly 3400 without SME. The final version of Lunar Lake will score nearly 3000-3100 in Linux.

Yes, Geekbench is a platform-dependent benchmark. Scores in Linux are 5-10% higher than in Windows, even while running in WSL. That's another reason to consider Geekbench as a garbage benchmark for cross-platform comparison (not to mention SME).

 

[Henry swagger]

Apple M3 scores nearly 3000-3100 in Geekbench, M4 - nearly 3400 without SME. The final version of Lunar Lake will score nearly 3000-3100 in Linux.

Yes, Geekbench is a platform-dependent benchmark. Scores in Linux are 5-10% higher than in Windows, even while running in WSL. That's another reason to consider Geekbench as a garbage benchmark for cross-platform comparison (not to mention SME).

Geekbench 5 is more valid for me than geekbench 6.. 5 correlates 1 to 1 with cibebench and specint17

 

[poke01]

Apple M3 scores nearly 3000-3100 in Geekbench, M4 - nearly 3400 without SME. The final version of Lunar Lake will score nearly 3000-3100 in Linux.

Yes, Geekbench is a platform-dependent benchmark. Scores in Linux are 5-10% higher than in Windows, even while running in WSL. That's another reason to consider Geekbench as a garbage benchmark for cross-platform comparison (not to mention SME).

So? It does so at a much higher clock. What part of that don’t you get? How can Lunar be more efficient than M3?

Also the M3 scores much higher than 3100. It does around ~3200.

 

[TwistedAndy]

Geekbench 5 is more valid for me than geekbench 6.. 5 correlates 1 to 1 with cibebench and specint17

Yes, Geekbench 5 is better, but there's still a 6% platform difference.

In general, SPEC is also not perfect, but there are not so many good options to choose from.

Also the M3 scores much higher than 3100. It does around ~3200.

Here are the latest Apple M3 results:

Spoiler: Geekbench Scores

Most of them are between 3000 and 3100. Yes, we can cherry-pick some 3200 or 2800 scores, but since we want to be objective, we take average scores.

 

[Magio]

Lunar Lake is expected to be more efficient than Apple M3 and nearly the same in terms of performance. It's expected to be an excellent platform for thin and light devices.

I really don't see anything to suggest that in the currently available scores. M3 scores over 10k in CB23 in the fanless Air, LNL needs 30W to get that scores and is 20+% below at 17W (which is probably still fanned). Single core is better but Intel has always drawn a lot more power single core vs Apple Silicon and we don't have any idea of what Lion Cove draws so let's wait on that front. If you want to speculate about final scores being a lot higher then that's your call but I'm commentating on what's currently out there, and those scores suggests worse efficiency at load and significantly worse performance.

Also, I've seen a few comments that LNL, with those scores, is still great for people who want a thin and light but don't want a Mac. That happens to be my case, I daily drive Linux so Apple hardware is a complete non-starter for me. But that doesn't change the fact that it empirically doesn't seem to compete with current Apple Silicon well at all. If I pay top of the line prices (which LNL will target), I expect top of the line products.

And yeah, LNL does still stand out because neither AMD nor Qualcomm have targeted the sub 20W power budget in this generation of products but in chipmaking you don't get a trophy just for showing up.

 

[poke01]

N

Yes, Geekbench 5 is better, but there's still a 6% platform difference.

In general, SPEC is also not perfect, but there are not so many good options to choose from.



Here are the latest Apple M3 results:

Spoiler: Geekbench Scores

View attachment 102022

Most of them are between 3000 and 3100. Yes, we can cherry-pick some 3200 or 2800 scores, but since we want to be objective, we take average scores.

No we take the highest when we compare performance. Taking the avg is silly when the M3 is also is used in fanless designs which will mess up averages.

Edit: looks only the M3 Pro and M3 Max due to having higher cache score in the 3200
Range. Looks like M3 is 3000-3150 range.

 

[TwistedAndy]

No we take the highest when we compare performance. Taking the avg is silly when the M3 is also is used in fanless designs which will mess up averages.

The highest and lowest values usually do not represent the actual value. For example, some measurements may be taken under extreme cooling. Another ones might be taken on the device having some heavy tasks running in the background. In both cases, the results are not representative.

That's why if we need to exclude the measurements outside a standard deviation, then take a mean value. It will be close to the performance you will actually get when you run a test.

I really don't see anything to suggest that in the currently available scores. M3 scores over 10k in CB23 in the fanless Air, LNL needs 30W to get that scores and is 20+% below at 17W (which is probably still fanned). Single core is better but Intel has always drawn a lot more power single core vs Apple Silicon and we don't have any idea of what Lion Cove draws so let's wait on that front. If you want to speculate about final scores being a lot higher then that's your call but I'm commentating on what's currently out there, and those scores suggests worse efficiency at load and significantly worse performance.

Also, I've seen a few comments that LNL, with those scores, is still great for people who want a thin and light but don't want a Mac. That happens to be my case, I daily drive Linux so Apple hardware is a complete non-starter for me. But that doesn't change the fact that it empirically doesn't seem to compete with current Apple Silicon well at all. If I pay top of the line prices (which LNL will target), I expect top of the line products.

And yeah, LNL does still stand out because neither AMD nor Qualcomm have targeted the sub 20W power budget in this generation of products but in chipmaking you don't get a trophy just for showing up.

Apple M3 consumes nearly ~30W in CB R23 MT and then throttles to ~20W. As for ST load, the P-core in M3 consumes nearly 6-7W under full load. Here are the actual power charts for Apple MacBook Air 13 with M3:


According to Geekerwan, the P-core in M4 will consume nearly 9W.

I don't think Lunar Lake will consume more power than M3. Laptop vendors will probably decide to go with the default 17-28W range for PL1. There's no need to go above it because Intel will release the ARL-H and ARL-HX in January.

 

[Magio]

Apple M3 consumes nearly ~30W in CB R23 MT and then throttles to ~20W. As for ST load, the P-core in M3 consumes nearly 6-7W under full load. Here are the actual power charts for Apple MacBook Air 13 with M3:
View attachment 102029

According to Geekerwan, the P-core in M4 will consume nearly 9W.

I don't think Lunar Lake will consume more power than M3. Laptop vendors will probably decide to go with the default 17-28W range for PL1. There's no need to go above it because Intel will release the ARL-H and ARL-HX in January.

It might not outright consume more but it will have comparable peak power draw (all LNL SKUs have a 30W PL2) and significantly lower performance across power profiles. The 17W PL1 benchmarks should be a decent approximation of the M3 in terms of power profile (short burst at 30W for both, dropping to 17/20W sustained) but across the board LNL delivers 20/30% worse performance. At 30W sustained, LNL barely matches M3 in Cinebench.

Apple Silicon single core power draw has been trending up, that's true, but IIRC on the most recent U series chips Intel's single core still was measurably more power hungry, drawing well over 10W under full load. Maybe Lion Cove changes that, but I'd be surprised if it had lower consumption than M3 or M4.

 

[majord]

I really don't see anything to suggest that in the currently available scores. M3 scores over 10k in CB23 in the fanless Air, LNL needs 30W to get that scores and is 20+% below at 17W (which is probably still fanned). Single core is better but Intel has always drawn a lot more power single core vs Apple Silicon and we don't have any idea of what Lion Cove draws so let's wait on that front. If you want to speculate about final scores being a lot higher then that's your call but I'm commentating on what's currently out there, and those scores suggests worse efficiency at load and significantly worse performance.

Also, I've seen a few comments that LNL, with those scores, is still great for people who want a thin and light but don't want a Mac. That happens to be my case, I daily drive Linux so Apple hardware is a complete non-starter for me. But that doesn't change the fact that it empirically doesn't seem to compete with current Apple Silicon well at all. If I pay top of the line prices (which LNL will target), I expect top of the line products.

And yeah, LNL does still stand out because neither AMD nor Qualcomm have targeted the sub 20W power budget in this generation of products but in chipmaking you don't get a trophy just for showing up.

Haven't targeted it explicitly , but even an old 6850U can pull 7100 CB23 @ 12w , 9300 @ 19w package power. It's perf/w peaked around that 12w score mark. No reason to think the 8+4c Zen 5 products won't have similar characteristics .

and That was a 2 yr old 8c product, no 'c' cores in sight, LnL should be playing in that same ballpark if it wants to be competitive. Would be a big worry if it was.