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 (20A)	Arrow Lake (N3B)	Lunar Lake	Panther Lake
Platform	Mobile H/U Only	Desktop Only	Desktop & Mobile H&HX	Mobile U Only	Mobile H
Process Node	Intel 4	Intel 20A	TSMC N3B	TSMC N3B	Intel 18A
Date	Q4 2023	Q1 2025 ?	Desktop-Q4-2024 H&HX-Q1-2025	Q4 2024	Q1 2026 ?
Full Die	6P + 8P	6P + 8E ?	8P + 16E	4P + 4E	4P + 8E
LLC	24 MB	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.)

 

[naukkis]

No one said Intel will be a leader but it'll be much closer this time, within 10%. That's the whole point. SPR is their weakest point in servers.

Intel is all in, two big dies glued together with interposer - thats pretty much all they could do without minding expenses and they are still losing to Amd cheap small chiplet products. Intel is so deep in shit they should sell whole server division to save rest of them.

 

[DokiDoki]

Within 10% in AVX-512 workloads too?

I think Intel saw that threat incoming and embarked on the AMX bandwagon. At least on that front, they have nothing to worry about.

What threat? Intel dictates the path that x86 vector-ISA will follow.

 

[511]

What threat? Intel dictates the path that x86 vector-ISA will follow.

This AMDs dominance is due to SW stack matured by Intel and the community if MI300 could run cuda nvidia wouldn't have a moat as Big

 

[DokiDoki]

This AMDs dominance is due to SW stack matured by Intel and the community if MI300 could run cuda nvidia wouldn't have a moat as Big

There's more to it than simply being able to 'run' CUDA. NVIDIA literally set the larger trends that were ultimately adopted in GPGPU computing.

Their current success is a continuation of that.
Why else would OpenCL be nonexistent today?

An analogy with Raytracing in video games: RT games are dime a dozen these days. They are all based on DX12. Because it's being pushed by Microsoft.

There is only one RT game that isn't DX12, and that is Doom Eternal.

 

[511]

There's more to it than simply being able to 'run' CUDA. NVIDIA literally set the larger trends that were ultimately adopted in GPGPU computing.

Yeah

 

[jdubs03]

Early UK pricing.

https://videocardz.com/newz/intel-core-ultra-9-285k-7-265kf-and-5-245kf-listed-by-first-retailer

 

[SiliconFly]

Apple sells about 15-20% of worldwide unit sales. Profits are much higher. The opposite with Intel, server marketshare higher and profits lower when compared to the competition. Not great.

In the global smartphone industry. Apple has about 15-20% marketshare in terms of volume shipped.

But in terms of profit, they take home 70-80% of all the profit in the industry. Android OEMs have to fight for the scraps.

Yeah. Thats Apple. But are you trying to compare Apple with AMD?

People get real nervous when there are any chances of good competition between the two. I though that was good for consumers.

After many many years, competition is all set to be hot this year. In clients, Lion Cove + Skymont has already surpassed Zen 5 (by a small margin). In servers, Granite Rapids is all set to catch up with competition. Interesting times ahead.

 

[jdubs03]

On that Royal thread I mentioned the other day. Looks like someone from that team posted in there. First of all what a name lmao. And two in my read they seem to poo poo it a bit (the conversation has more above and below).

 

[SiliconFly]

Intel is all in, two big dies glued together with interposer - thats pretty much all they could do without minding expenses and they are still losing to Amd cheap small chiplet products. Intel is so deep in shit they should sell whole server division to save rest of them.

I don't think so! GNR does not use interposers afaik. The interposers will be so large which would make the whole SoC unviable. Instead, the chiplets reside directly on the substrate and are glued together using EMIB (like TSMC InFO).

 

[igor_kavinski]

What threat? Intel dictates the path that x86 vector-ISA will follow.

Haven't seen the AVX-512 benchmarks for Zen 5, have you? Or the fact that Zen 5 is doing AVX-512 workloads better than Intel without consuming extra power or downclocking like crazy?

Yes, Intel will probably shift the goal with AVX10 but other than introducing few new instructions, they can't do anything to stop legacy AVX-512 workloads from running on Ryzens or Turin Epycs with better thermals and power usage, making them super attractive to anyone depending on these workloads.

 

[DokiDoki]

Haven't seen the AVX-512 benchmarks for Zen 5, have you? Or the fact that Zen 5 is doing AVX-512 workloads better than Intel without consuming extra power or downclocking like crazy?

Yes, Intel will probably shift the goal with AVX10 but other than introducing few new instructions, they can't do anything to stop legacy AVX-512 workloads from running on Ryzens or Turin Epycs with better thermals and power usage, making them super attractive to anyone depending on these workloads.

You're talking as if Xeons cannot run AVX512. And downclocking with AVX-512 has been solved on Intel since Ice Lake.

 

[igor_kavinski]

You're talking as if Xeons cannot run AVX512.

OpenSSL Benchmark - OpenBenchmarking.org

openbenchmarking.org

Why isn't Intel on top there, since they are supposed to be the original inventor of AVX-512?

 

[511]

Haven't seen the AVX-512 benchmarks for Zen 5, have you? Or the fact that Zen 5 is doing AVX-512 workloads better than Intel without consuming extra power or downclocking like crazy?

Yes, Intel will probably shift the goal with AVX10 but other than introducing few new instructions, they can't do anything to stop legacy AVX-512 workloads from running on Ryzens or Turin Epycs with better thermals and power usage, making them super attractive to anyone depending on these workloads.

The same AVX that AMD Copied they may have better hardware i can bet half of AVX-512 software is intel developed software 🤣 your hardware is sand if it doesn't get supported intel will just use underhanded technique to block them from using intel software they have done it in the past

 

[DokiDoki]

OpenSSL Benchmark - OpenBenchmarking.org

openbenchmarking.org

Why isn't Intel on top there, since they are supposed to be the original inventor of AVX-512?

?

Why is the most powerful automobile ICE made by Bugatti when the first to put the ICE in a car was Diamler?

 

[igor_kavinski]

your hardware is sand if it doesn't get supported intel will just use underhanded technique to block them from using intel software they have done it in the past

Don't know a single case of that happening so far. Do enlighten me if there is some AVX-512 software using the same instructions available on AMD hardware but still won't run.

 

[igor_kavinski]

?

Why is the most powerful automobile ICE made by Bugatti when the first to put the ICE in a car was Diamler?

And we can do this all day long then. Point is, Intel lagged behind. They ARE in trouble. Now they don't have enough resources to do both server and client. So what are they gonna choose?

 

[511]

Don't know a single case of that happening so far. Do enlighten me if there is some AVX-512 software using the same instructions available on AMD hardware but still won't run.

I don't know either 🤣 i will wait for them to pull it you can't change the underhanded technique intel nvidia had used over the years

 

[jdubs03]

And we can do this all day long then. Point is, Intel lagged behind. They ARE in trouble. Now they don't have enough resources to do both server and client. So what are they gonna choose?

Ehh tbh I think their lineup for next year with Panther Lake and Clearwater Forest/Diamond Rapids are pretty solid. Much much more compelling on the server side than anything on offer now (including 6th gen Xeon).

 

[igor_kavinski]

Ehh tbh I think their lineup for next year with Panther Lake and Clearwater Forest/Diamond Rapids are pretty solid.

That's great if they are still here next year!

 

[511]

That's great if they are still here next year!

AMD has survived worst tbh

 

[igor_kavinski]

AMD has survived worst tbh

AMD wasn't full of absolute morons working against each other.

Intel's teams and management are used to having too much resources at their disposal. AMD has ALWAYS been lean. They don't know what excess is.

Going from having anything available to limited resources doesn't immediately lead to better execution. It will actually cause a lot of teething pains as they discover how to do things with limited budgets and those unable to thrive in such an environment will quit and leave gaping holes in skills and expertise.

Their stupid Meteor Lake hot lots fiasco is a direct result of them thinking they can do anything with their resources, in other words, waste investor money!

 

[SiliconFly]

That's great if they (intel) are still here next year!

AMD has survived worst tbh

AMD wasn't full of absolute morons working against each other.

Is this truth time?

Ehh tbh I think their lineup for next year with Panther Lake and Clearwater Forest/Diamond Rapids are pretty solid. Much much more compelling on the server side than anything on offer now (including 6th gen Xeon).

And we can do this all day long then. Point is, Intel lagged behind. They ARE in trouble. Now they don't have enough resources to do both server and client. So what are they gonna choose?

I don't think the client/server groups are under any threat (at least in the short term). PNL/CWF/DNR are all on track and are very promising! If they deliver as promised, competition is in very serious trouble considering their already minuscule market share both in clients AND servers!

 

[jdubs03]

AMD wasn't full of absolute morons working against each other.

Intel's teams and management are used to having too much resources at their disposal. AMD has ALWAYS been lean. They don't know what excess is.

Going from having anything available to limited resources doesn't immediately lead to better execution. It will actually cause a lot of teething pains as they discover how to do things with limited budgets and those unable to thrive in such an environment will quit and leave gaping holes in skills and expertise.

Such a shame because on the engineering side they’re quite competent.

 

[DokiDoki]

They ARE in trouble. Now they don't have enough resources to do both server and client.

According to who?

 

[igor_kavinski]

Such a shame because on the engineering side they’re quite competent.

Yeah but engineering being overridden by dumb people who think that the average person buys TWO expensive laptops, one as a backup!