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.)

 

[Geddagod]

It's aggregate performance. Pure speculation based on frequency to core count among other factors. It's as good as pulling random stats out of your behind. It's used professionally for a really jank estimate of what something may perform like. If the figures were remotely true Intel's desktop would be dead in the water again for years to come given what AMD wants to do.

Aggregate performance is pretty much just pure MT performance I'm pretty sure. That's why Intel also said, in the same files, that SRF also targets high aggregate performance while GNR is targeted towards tasks that have latency considerations, etc etc.
What do these figures have to do with Intel's desktop lineup vs what AMD wants to do? What "aggregate performance" claims did Intel make on desktop? I'm so confused....
I mean it is an internal target for 2X MT performance, so of course no one is saying it will happen officially, but why do you think it is a "really jank estimate"?
And you kinda seem to contradict yourself literarily in the first two sentences. If you are basing your performance target on core counts and frequency and IPC and other factors such as core scaling, etc etc, then it seems like at worst it's an educated guess, not a random pull from your behind.

 

[Geddagod]

The last leak posted by YuuKi_Ans would indicate a 12-15% performance boost vs Raptor Cove, I suspect that is 10% IPC + 5 speed on Lion Cove cores(being conserbative, it could 8 IPC 7% core speed)

View attachment 78776

Also forgot to add, why talk about Lion Cove cores about MTL? MTL uses redwood cove cores. Also GNR might have been planned to use Lion Cove at one point, which I doubt but seems to be what Exist50 thinks, but GNR is also supposed to use RWC or RWC+.
And in the very same leak you are attaching on your post, it says Redwood Cove core right there so.... ???

 

[A///]

Aggregate performance is pretty much just pure MT performance I'm pretty sure. That's why Intel also said, in the same files, that SRF also targets high aggregate performance while GNR is targeted towards tasks that have latency considerations, etc etc.
What do these figures have to do with Intel's desktop lineup vs what AMD wants to do? What "aggregate performance" claims did Intel make on desktop? I'm so confused....
I mean it is an internal target for 2X MT performance, so of course no one is saying it will happen officially, but why do you think it is a "really jank estimate"?
And you kinda seem to contradict yourself literarily in the first two sentences. If you are basing your performance target on core counts and frequency and IPC and other factors such as core scaling, etc etc, then it seems like at worst it's an educated guess, not a random pull from your behind.

Only in the most basic way, it's really freq times the number of available physical cores. It accounts for nothing else including underlying architecture or any hardware accelerators which intel's new xeons have and will continue to gain. AMD has the same for ryzen and epyc/threadripper.

It won't take into consideration anything but the frequency and core count. That is it. It's a very basic guideline to go through. Take everything into consideration and going by the aggregate your estimate to irw performance may differ anywhere from 5%-40%, sometimes higher. It would make sense in his example leak if the cores for p + e were roughly the same with some minor upgrades and no changes to the cache structure of the processor or refinement in the pipeline. Basically anything.

 

[Geddagod]

Only in the most basic way, it's really freq times the number of available physical cores. It accounts for nothing else including underlying architecture or any hardware accelerators which intel's new xeons have and will continue to gain. AMD has the same for ryzen and epyc/threadripper.

Really? Doesn't account for IPC or fabric or anything else?
Is this some hardware term that's standardized throughout the industry or something Intel and AMD has done in the past? Because I have never heard of it, but it would be nice if you could show ma en example.
I don't mind the performance metric not including dedicated hardware accelerators because that could be highly variable.

 

[Geddagod]

If Intel is going to see a significant increase in battery life, it will come with Lunar Lake or not at all. I doubt Meteor Lake will provide a significant improvement.

Low power SOC island has potential IMO if they can handle basic web surfing while essentially just turning off the entire CPU tile.

 

[A///]

Really? Doesn't account for IPC or fabric or anything else?
Is this some hardware term that's standardized throughout the industry or something Intel and AMD has done in the past? Because I have never heard of it, but it would be nice if you could show ma en example.
I don't mind the performance metric not including dedicated hardware accelerators because that could be highly variable.

Not in the non x86 world myself. We've always counted as base frequency times the number of available cores. There's obvious flaws in this concept. One of which anyone is familiar with was the frequency regression going from hot pentium 4 to the core lineup, except the latter performed the same or higher than the outgoing pentium d and pentium 4s while running cooler, and had a lot of breathing room for overclocking.

With amd because I can't be bothered to learn Intel's trash lineup is that you can take a 3600x and a 3950x, times their base clock speed to the number of cores, and find the percentage difference and then you can say the 3950X is x per cent faster. That's not really how it works, especially if you compared a 1600x to a 5600x. at its most simplest form you're comparing on paper what the processors may be capable of doing. In reality once you go cross generational there's too much to consider. It's why you'll see a lower core 12 th gen smack around any skylake based processor with higher core counts with the latter skylake only gaining some ground in processing intensive tasks like video encoding which may have a marginal or small sizeable win.

You can find better explained reasoning online if you were to look up why no one takes aggregate performance as fact. It's good for off the cuff assessment, but nothing I'd use professionally. You can see the flaw of the method earlier in Ryzen gen 1000 going up against intel. high core count Ryzens just matching slightly lower count Intels. The aggregate numbers favor Ryzen then but in reality intel still pulled ahead.

 

[Geddagod]

Not in the non x86 world myself. We've always counted as base frequency times the number of available cores. There's obvious flaws in this concept. One of which anyone is familiar with was the frequency regression going from hot pentium 4 to the core lineup, except the latter performed the same or higher than the outgoing pentium d and pentium 4s while running cooler, and had a lot of breathing room for overclocking.

With amd because I can't be bothered to learn Intel's trash lineup is that you can take a 3600x and a 3950x, times their base clock speed to the number of cores, and find the percentage difference and then you can say the 3950X is x per cent faster. That's not really how it works, especially if you compared a 1600x to a 5600x. at its most simplest form you're comparing on paper what the processors may be capable of doing. In reality once you go cross generational there's too much to consider. It's why you'll see a lower core 12 th gen smack around any skylake based processor with higher core counts with the latter skylake only gaining some ground in processing intensive tasks like video encoding which may have a marginal or small sizeable win.

You can find better explained reasoning online if you were to look up why no one takes aggregate performance as fact. It's good for off the cuff assessment, but nothing I'd use professionally. You can see the flaw of the method earlier in Ryzen gen 1000 going up against intel. high core count Ryzens just matching slightly lower count Intels. The aggregate numbers favor Ryzen then but in reality intel still pulled ahead.

I'm not trying to argue that calculating performance like that is wrong, but rather that's what Intel did when they said aggregate performance. I would think since that it's internal targets, they would at the very least also include IPC in their predictions, but idk tbh.
Either way though it seems like if aggregate only factors in clocks and core counts, then GNR might be greater than 2X expected because of slightly higher IPC and better bandwidth per core...
Atleast in tasks that are embarrassingly parallel (like I'm pretty sure CB and GB 5 are). Unless mesh faces serious bandwidth issues, which who knows could happen, despite that being a focus on the platform.

 

[Markfw]

Literally on the leak you were quoting, it says Granite Rapids AP, targeting ~2X performance of EMR-64C.
And also I'm pretty sure, at the very least, 2X SPR is enough to beat Genoa.
SPR 56C scores ~65K on CBR23 MT while Genoa scores 115K apparently.
In GB5, 2X the 56C SPR would be 108% of top end Genoa as well.
These are the only MT scores I can find comparing the two. If you can find some that show otherwise, I would be happy to hear them....

The key word there is TARGETING. So the best they came come up with is a TARGET of 2x a currently available CPU

I will believe it when I see hard facts and benchmarks to prove it. In the meantime, Intel is hurting.... THAT IS FACT.

 

[A///]

I'm not trying to argue that calculating performance like that is wrong, but rather that's what Intel did when they said aggregate performance. I would think since that it's internal targets, they would at the very least also include IPC in their predictions, but idk tbh.
Either way though it seems like if aggregate only factors in clocks and core counts, then GNR might be greater than 2X expected because of slightly higher IPC and better bandwidth per core...
Atleast in tasks that are embarrassingly parallel (like I'm pretty sure CB and GB 5 are). Unless mesh faces serious bandwidth issues, which who knows could happen, despite that being a focus on the platform.

Honestly to tell you the truth here I don't think even Intel knows their final numbers. They're so hard into crunch mode to get these out that I would not be surprised if their Xeons turn out to be security flawed processors a decade from now or need to be recalled at some point. To them delivery is delivery and they can fix it down the line. There's a lot about Intel in the next 3-5 years that are so uncertain that it was 2017 and I was playing with Zen and wondering how long Intel would take to give AMD an answer. I'm afraid Rocket, Alder and Raptor aren't quite the answer I had in mind. I cheerlead Gelsinger because I really do want Intel to make a comeback. It benefits the consumer. This is one of the few times I bite my tongue not wanting to say what I truly think about Intel's future as if my words become their future roadpath to failure. This is why I joked not long ago that Intel is silently praying Zen 5 is a nothingburger.

 

[Geddagod]

The key word there is TARGETING. So the best they came come up with is a TARGET of 2x a currently available CPU

I will believe it when I see hard facts and benchmarks to prove it. In the meantime, Intel is hurting.... THAT IS FACT.

 

[Geddagod]

Honestly to tell you the truth here I don't think even Intel knows their final numbers. They're so hard into crunch mode to get these out that I would not be surprised if their Xeons turn out to be security flawed processors a decade from now or need to be recalled at some point. To them delivery is delivery and they can fix it down the line. There's a lot about Intel in the next 3-5 years that are so uncertain that it was 2017 and I was playing with Zen and wondering how long Intel would take to give AMD an answer. I'm afraid Rocket, Alder and Raptor aren't quite the answer I had in mind. I cheerlead Gelsinger because I really do want Intel to make a comeback. It benefits the consumer. This is one of the few times I bite my tongue not wanting to say what I truly think about Intel's future as if my words become their future roadpath to failure. This is why I joked not long ago that Intel is silently praying Zen 5 is a nothingburger.

Ye I don't think Intel knows their numbers exactly, and even if they did there is no certainty they would become true. Completely agree.

 

[Markfw]

Ye I don't think Intel knows their numbers exactly, and even if they did there is no certainty they would become true. Completely agree.

OK, just to get it straight, we all agree that Intel is an unknown , they are hurting now, and their future is a complete unknown, just a wish list.

 

[Geddagod]

OK, just to get it straight, we all agree that Intel is an unknown , they are hurting now, and their future is a complete unknown, just a wish list.

Sure. No one here said otherwise. But lets not move goalposts.
Ok, just to get it straight, we all agree that this leak claimed Granite Rapids would compete with Genoa

 

[A///]

Ye I don't think Intel knows their numbers exactly, and even if they did there is no certainty they would become true. Completely agree.

I refer you to the 80s big hair heavyset youtuber who made a mockery of intel's new mobile lineup a few years ago because they wouldn't shut up about amd or what their product actually stood for. not the actions of a confident company.

 

[A///]

OK, just to get it straight, we all agree that Intel is an unknown , they are hurting now, and their future is a complete unknown, just a wish list.

I can just about imagine Pat "Guns A'Blazing Gunslinger" Gelsinger sitting in the corner of his dark office, drink in hand listening to Foreigner's album anthologies crying himself to sleep.

 

[Markfw]

Sure. No one here said otherwise. But lets not move goalposts.
Ok, just to get it straight, we all agree that this leak claimed Granite Rapids would compete with Genoa

"claimed" being the important word. Which I have NO confidence in. Like if "claiming" that Genoa is 5 times faster than SPR. We all know thats crap.

lets stick to something close to reality.

 

[mderbarimdiger]

markfw, this thread is designated for speculations and discussions about future chips, if you prefer to stick close to reality and you want to see benchmarks please check out raptor lake and Zen 4 threads.

rest assured that everyone here understood that a leak about a target from intel about future potential chips is not considered a fact, and rest assured, Xeon is not in the same class as Genoa, mostly everyone is well aware of that fact.

 

[Markfw]

markfw, this thread is designated for speculations and discussions about future chips, if you prefer to stick close to reality and you want to see benchmarks please check out raptor lake and Zen 4 threads.

rest assured that everyone here understood that a leak about a target from intel about future potential chips is not considered a fact, and rest assured, Xeon is not in the same class as Genoa, mostly everyone is well aware of that fact.

And why is it that you felt the need to post this ? I don't believe that a FUTURE chip is what they are saying. That is ON TOPIC. You are a relatively new poster and you felt the need to insult my intelligence with this post ?

 

[A///]

there is speculation and then there is talking rubbish. Let's not all fall into the lands of my life is dismal and other social pariahs masquerading as technocrats with insider knowledge.

 

[ashFTW]

And why is it that you felt the need to post this ? I don't believe that a FUTURE chip is what they are saying. That is ON TOPIC. You are a relatively new poster and you felt the need to insult my intelligence with this post ?

Intimidation of new posters, eh?

 

[Henry swagger]

Intimidation of new posters, eh?

How is that guy a mod 😯😁 ?



Mod callouts are not allowed.


esquared
Anandtech Forum Director

 

[Henry swagger]

markfw, this thread is designated for speculations and discussions about future chips, if you prefer to stick close to reality and you want to see benchmarks please check out raptor lake and Zen 4 threads.

rest assured that everyone here understood that a leak about a target from intel about future potential chips is not considered a fact, and rest assured, Xeon is not in the same class as Genoa, mostly everyone is well aware of that fact.

Stop it

 

[coercitiv]

You are a relatively new poster and you felt the need to insult my intelligence with this post ?

How old does a forum poster need to be before you understand this endless quarrel is completely detrimental to the health of the forum? I'm tired of the same noise every day, I read less and interact less because people actively choose to inject their opinions everywhere between the few actual good leaks we get on new products. For me the highlight of the Intel threads was those Granite Rapids and Sierra Forest slides, I finally got to read things without constant interjections from overzealous forum members.

Constant arguments drive the typical users away, and you will be left with partizans willing to stomach the trolling.

 

[ashFTW]

How old does a forum poster need to be before you understand this endless quarrel is completely detrimental to the health of the forum? I'm tired of the same noise every day, I read less and interact less because people actively choose to inject their opinions everywhere between the few actual good leaks we get on new products. For me the highlight of the Intel threads was those Granite Rapids and Sierra Forest slides, I finally got to read things without constant interjections from overzealous forum members.

Constant arguments drive the typical users away, and you will be left with partizans willing to stomach the trolling.

I support this opinion @coercitiv, I have the same reaction to these kinds of constant posts. I’m mostly vendor agnostic. AMD has made me a lot of money in the past, and my current workstation (which I absolutely love) is a Threadripper, but I also enjoy thinking and speculating about chip architectures from all other vendors including Intel and Apple. I am very close to leaving this forum; it’s become quite toxic!

 

[mikk]

And why is it that you felt the need to post this ? I don't believe that a FUTURE chip is what they are saying. That is ON TOPIC. You are a relatively new poster and you felt the need to insult my intelligence with this post ?

It is very important for you because you are clearly not aware this is a speculation thread and not everything can be backed up with facts. He wants to help you explaining it, it's not a bad thing.