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.)

Thunder 57 · Jan 16, 2025

OneEng2 said:
Lets be clear, Apple nearly went under trying to compete with x86 PC's using ARM. The invention of the iPOD and subsequent creation of the iPhone saved the company ...... it WASN'T ARM.

Pretty sure you mean POWER .

poke01 · Jan 16, 2025

OneEng2 said:
, Apple nearly went under trying to compete with x86 PC's using ARM

What? They never used ARM on Macs before 2020

MoistOintment · Jan 16, 2025

OneEng2 said:
Lets be clear, Apple nearly went under trying to compete with x86 PC's using ARM. The invention of the iPOD and subsequent creation of the iPhone saved the company ...... it WASN'T ARM.

I thought it was clear I was talking about the Apple M series and not literal decades ago. Apple moved 10% of the client PC market onto ARM with that decision. That's a 10% TAM shrink for one of Intel's most important segments.

The x86 doom and gloom is really obvious when you look at trends and trajectories. I'm not going to put a firm timeline or prediction on dates - but I do think it's clear that x86 will not be the dominate ISA in client and server within my lifetime. If the client PC segment becomes majority ARM, then that severely raises costs for x86 development for server and increasingly motivates a transition to ARM.

This will be a multi-decade long transition. It won't be immediate. It won't be soon. But it's literally playing out right now.
Nvidia will bring more of their server solutions over to ARM. Their partnership with MediaTek is only the beginning.

The whole point of this doom-and-gloom is to point out that Intel (and AMD) need to focus on alternatives to x86 - and they are. dGPUs and for Intel, also their Fabs.

Hulk · Jan 16, 2025

OneEng2 said:
Lets be clear, Apple nearly went under trying to compete with x86 PC's using ARM. The invention of the iPOD and subsequent creation of the iPhone saved the company ...... it WASN'T ARM.

Truer words were never spoken.

oak8292 · Jan 17, 2025

OneEng2 said:
Interesting POV that people would be OK with a 50-70% performance decrease. Hmmm.

Lets be clear, Apple nearly went under trying to compete with x86 PC's using ARM. The invention of the iPOD and subsequent creation of the iPhone saved the company ...... it WASN'T ARM.

ARM processors make very good low power devices. No argument. I have designed products around ARM, Coldfire, 68332, HC12, the ill fated Intel Automotive SOC, PIC, and yes, x86 for embedded devices.

For the higher processor requirements jobs, x86 worked better. For true embedded, or even linux based embedded, ARM usually fits the bill. For REALLY small things that need to be very inexpensive, draw little power, etc, I don't even bother with Linux, but rather with FreeRTOS that is stripped down to the elements I need for the product. In this last case (see your example of VR headset), you want DSP doing all the work and the RTOS is pretty darned slim. It doesn't matter if you are using ARM or some other ISA for these applications as the hardware and its built in capabilities is much more important than just the ISA.

For server work, x86 is currently untouchable with the exception of specific use cases by VERY large companies that can afford to custom make their entire software stack to be optimized for their own ARM design cpu's and platforms.

It's crazy all this "x86 doom, doom doom on you" talk in here!

Here is an excerpt from AWS re:Invent transcript about Amazon’s fourth generation ARM server processor.

“Now in terms of customer momentum, we see customers across all categories, geographies,
10:19
verticals, and all sizes adopting Graviton in a big way.
10:24
And today 95% of the top thousand EC2 customers are drawing on Graviton.
10:34
That's not all, and this was just announced earlier in re:Invent. Over the last two years,
10:39
more than 50% of new compute capacity that we've deployed is based on AWS Graviton.”

Over the last two years “more than 50% of new compute capacity that we’ve deployed is based on AWS Graviton. The software moat is getting smaller when the largest cloud vendor is using that much ARM.

There are still customers for IBM Z systems with Power PC ISA. There will continue to be customers for x86 for generations but there is also a lot of room for ARM and accelerators. Eventually it could be RISC-V with zero royalty and just the cost of engineering architecture.

Computers are powerful enough for emulation and abstraction for many jobs. Custom workloads will be on accelerators (GPUs) or custom hardware and that may be an ASIC (crypto) if the workload is large enough.

okoroezenwa · Jan 17, 2025

Hulk said:
Truer words were never spoken.

Even though some of it is not true?

SiliconFly · Jan 17, 2025

Hulk said:
Truer words were never spoken.

okoroezenwa said:
Even though some of it is not true?

True! 🤔

511 · Jan 17, 2025

Intel 3 vs Intel 4 10% performance from a simple port

LightningZ71 · Jan 17, 2025

That moves that product firmly out of the 12th gen Alder Lake U range with respect to performance finally. It's likely around 1345/1355U range in single thread, maybe a bit better in multi.

dullard · Jan 17, 2025

511 said:
Intel 3 vs Intel 4 10% performance from a simple port

I don't see why you posted about Intel in this ARM/Apple/Linux (and now POWER) thread.

dullard · Jan 17, 2025

511 said:
Intel 3 vs Intel 4 10% performance from a simple port

Seriously though, 10% performance gain isn't much. But 10% performance gain at just about half the power is quite an improvement.

511 · Jan 17, 2025

dullard said:
I don't see why you posted about Intel in this ARM/Apple/Linux (and now POWER) thread.

Nooooo Wrong thread again I would like an exclusive thread where whatever I say is correct
/s

511 · Jan 17, 2025

dullard said:
Seriously though, 10% performance gain isn't much. But 10% performance gain at just about half the power is quite an improvement.

We still don't know power but I doubt it's half we need someone to do testing but at same performance if it is consuming 20-30% less power than it is very nice improvement

SiliconFly · Jan 17, 2025

Looks like Intel has finally managed to release the so called performance updates for Arrow Lake.

Intel claims Core Ultra 200 patches improve gaming performance by up to 26%

Intel quickly investigated the situation, and by December, it had a good idea of what was causing the performance hiccups. It turned out to be a combination...

www.techspot.com

SiliconFly · Jan 17, 2025

Did any of you follow the interesting rumors about Intel acquisition? It's all over X/Twitter and it appears semiaccurate site crashed for a while due to excessive traffic.

Hulk · Jan 17, 2025

511 said:
Intel 3 vs Intel 4 10% performance from a simple port View attachment 115115

This is what used to be a "Tick."
Specs and score are nice but to complete the picture we need the average frequency and package power during the run to make a valid comparison. One system could simply have better cooling and be boosting higher.

dullard · Jan 17, 2025

511 said:
We still don't know power but I doubt it's half we need someone to do testing but at same performance if it is consuming 20-30% less power than it is very nice improvement

I was just going off the table's listed power. But, your point is correct that the table doesn't capture the exact power level actually consumed.

511 · Jan 17, 2025

Hulk said:
This is what used to be a "Tick."
Specs and score are nice but to complete the picture we need the average frequency and package power during the run to make a valid comparison. One system could simply have better cooling and be boosting higher.

I agree we need someone to do the testing so if true Intel 3 will be indeed a full node worth of PPW

Thunder 57 · Jan 17, 2025

SiliconFly said:
Looks like Intel has finally managed to release the so called performance updates for Arrow Lake.

Intel claims Core Ultra 200 patches improve gaming performance by up to 26%

Intel quickly investigated the situation, and by December, it had a good idea of what was causing the performance hiccups. It turned out to be a combination...

www.techspot.com

I'd wait for independent verification. Even the Intel slide says single digit gain in a geomean of 35 games. Performances fixs have been claimed before and never seem to work out.

It reminds me of Bulldozer. "That damn Windows scheduler! It'll be patched and then Windows 8 will fix it once and for all!". Never amounted to anything really.

DavidC1 · Jan 17, 2025

The everlasting x86 vs ARM argument still depends on whether the x86 duopolies can actually continue to execute.

Things like saying Lunarlake is a one-off doesn't inspire confidence, but they still have time. It's clear the transition will take long time, but depending on what the x86 vendors do it can be accelerated.

Doug S · Jan 17, 2025

OneEng2 said:
Interesting POV that people would be OK with a 50-70% performance decrease. Hmmm.

* 30-50% decrease, because I said 50-70% of native.

Lets call it 50% performance, you run x86 apps on your Parallels ARM Windows install and you lose half the performance of the M4. Why do you think that would be a problem? The M4 is faster than any x86, so the decrease over the x86 you might buy instead would be less than 50%. The decrease over the x86 you are REPLACING would be even less. Perhaps not even noticeable.

Look no one is going to buy an M4 Mac if all they do every day is run Windows apps that are only available on x86. You'd have to be brain damaged to think that makes sense. If you have a need to run some Windows stuff part time, and you're unlucky that there is no ARM port of that Windows stuff, then you'd take a performance hit. But so what? That hit only matters if your "Windows stuff" is performance sensitive, and MOST software is not performance sensitive.

SiliconFly · Jan 17, 2025

Doug S said:
* 30-50% decrease, because I said 50-70% of native.

Lets call it 50% performance, you run x86 apps on your Parallels ARM Windows install and you lose half the performance of the M4. Why do you think that would be a problem? The M4 is faster than any x86, so the decrease over the x86 you might buy instead would be less than 50%. The decrease over the x86 you are REPLACING would be even less. Perhaps not even noticeable.

Look no one is going to buy an M4 Mac if all they do every day is run Windows apps that are only available on x86. You'd have to be brain damaged to think that makes sense. If you have a need to run some Windows stuff part time, and you're unlucky that there is no ARM port of that Windows stuff, then you'd take a performance hit. But so what? That hit only matters if your "Windows stuff" is performance sensitive, and MOST software is not performance sensitive.

First off, Parallels is a VM. Installing an app in a VM and running it is not for everyone. It's very resource intensive and the performance penalty is not just 30% less or 50% less, it's severe with heavy lag and stuttering and inefficient. Second is usability. For example, if the app saves a file, it is saved in the VM by default. People have to set up shares and use them. Not for everyone. Third, the VM also needs regular maintenance like Windows Updates, software updates, driver updates, etc. Not for everyone.

Even running Windows apps on a x64 VM on x64 Windows running on a x64 CPU is a pain. And you're talking about running a (incompatible) x64 VM on Mac OS running on a ARM CPU. It's just dumb imho. Parallels is more of a "desperate times, desperate measures" kinda thing.

Doug S · Jan 17, 2025

SiliconFly said:
First off, Parallels is a VM. Installing an app in a VM and running it is not for everyone. It's very resource intensive and the performance penalty is not just 30% less or 50% less, it's severe with heavy lag and stuttering and inefficient. Second is usability. For example, if the app saves a file, it is saved in the VM by default. People have to set up shares and use them. Not for everyone. Third, the VM also needs regular maintenance like Windows Updates, software updates, driver updates, etc. Not for everyone.

Even running Windows apps on a x64 VM on x64 Windows running on a x64 CPU is a pain. And you're talking about running a (incompatible) x64 VM on Mac OS running on a ARM CPU. It's just dumb imho. Parallels is more of a "desperate times, desperate measures" kinda thing.

I've never used Parallels but I used VMware for years to run a Windows VM under Linux and it worked great. No problems with "stuttering" and setting up shares to have stuff stored outside the VM is built in. It is not remotely as bad an experience are you are trying to imply.

DavidC1 · Jan 17, 2025

Translation doesn't guarantee compatibility, so there are still cases where it simply won't run at all. This is a bigger deal than the loss of performance.

SiliconFly · Jan 17, 2025

DavidC1 said:
Translation doesn't guarantee compatibility, so there are still cases where it simply won't run at all. This is a bigger deal than the loss of performance.

True. Also, the VM itself requires around 8GB to 10GB of RAM to run comfortably.

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

Senior member

Attachments

Diamond Member

Diamond Member

Member

Diamond Member

Member

Member

Golden Member

Golden Member

Platinum Member

Elite Member

Elite Member

Golden Member

Golden Member

Golden Member

Golden Member

Diamond Member

Elite Member

Golden Member

Diamond Member

Golden Member

Diamond Member

Golden Member

Diamond Member

Golden Member

Golden Member