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 (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.)

Abwx · Tuesday at 11:52 AM

511 said:
Granite rappids review out destroying amd genoa we are so back

Intel Xeon 6980P "Granite Rapids" Linux Benchmarks Review - Phoronix

www.phoronix.com

33% more cores and 39% higher power for a mere 11.8% better perf than a 2 years old CPU, quite a "win", if you set those 96C Epyc at the same 500W they would be on par, so if a 96C Zen 4 match Intel s 128C at same power i let you imagine what will happen soon with the Zen 5 based Epyc.

511 · Tuesday at 11:56 AM

Abwx said:
33% more cores and 25% higher power for a mere 11.8% better perf than a 2 years old CPU, quite a "win", if you set those 96C Epyc at the same 500W they would be on par, so if a 96C Zen 4 match Intel s 128C at same power i let you imagine what will happen soon with the Zen 5 based Epyc.

He didn't have power draw numbers lol from RAPL which he usually does but anyway it will still be better than 64C vs 96C from 2 years ago won't be a bloodbath it will be closer i still expect them to loose by 15-20% outside of their libraries

9949asd · Tuesday at 11:58 AM

Hitman928 said:
No, it boosts to 4.8 GHz. Maybe his laptop has a cooling issue or he’s mistaken on the clock speed.

It’s up to 4.8g, temperature over 75c -0.2g.

GTracing · Tuesday at 12:03 PM

Quick comparison of the gaming performance of Lunar Lake vs Meteor lake vs Strix Point. The second two are older reviews, so they likely score higher now. Power consumption is different between models as well. In Cyberpunk FHD ultra they consume 33.1W average, 40.7W average, and 49.2W average respectively. For ram they have 32GB LPDDR5x-8533, 32 GB LPDDR5x-7467, and 32 GB LPDDR5x-7500 respectively.

Overall good showing from Lunar Lake. Lunar lake pulls ahead of Meteor lake at higher graphics settings, but at lower settings it seems to have regressed in perf/watt in a couple games. I'm guessing the drivers need work.

Abwx · Tuesday at 12:05 PM

511 said:
He didn't have power draw numbers lol from RAPL which he usually does but anyway it will still be better than 64C vs 96C from 2 years ago won't be a bloodbath it will be closer i still expect them to loose by 15-20% outside of their libraries

There s a page about this subject here :

Welcome Back Intel Xeon 6900P Reasserts Intel Server Leadership

The Intel Xeon 6900P offers twice the cores, huge memory bandwidth and more connectivity. Welcome back Intel to the top end server CPU market

www.servethehome.com

Anyway based on these number we can infer that Zen 5 128C Epyc wil be largely ahead due to core count parity and also a slightly better process, add the augmented IPC and you get the picture

511 · Tuesday at 12:07 PM

Abwx said:
There s a page about this subject here :

Welcome Back Intel Xeon 6900P Reasserts Intel Server Leadership

The Intel Xeon 6900P offers twice the cores, huge memory bandwidth and more connectivity. Welcome back Intel to the top end server CPU market

www.servethehome.com

Anyway based on these number we can infer that Zen 5 128C Epyc wil be largely ahead due to core count parity and also a slightly better process, add the augmented IPC and you get the picture

There is no process advantage tbf this time Intel 3 is more or less on par with what Turin has there will not be largely this time the per core performance will be higher we will see in 16 days anyway

511 · Tuesday at 12:07 PM

Clearwater Forest

naukkis · Tuesday at 12:08 PM

MoogleW said:
月光下的新探索：Lunar Lake CPU (Lion Cove / Skymont) 微架构测试 | David Huang's Blog

That Skymont has better scores than Lion Cove in some microbenchmarks is rather impressive

Very interesting results. Skymont low-power performance is unseen before on x86. Even @3.8GHz there ain't power budget for 4 P-cores. P-cores are doomed, intel should have put only 2 p-cores and add 2 or 4 e-cores to the ring for better MT perf.

ikjadoon · Tuesday at 12:13 PM

CB23 1T tests, after SemiAccurate's claims (confirmed by ASUS to KitGuru)

Seems to be smaller than ASUS predicted. Perhaps some mitigation? The update itself also failed to update a few times, but finally went through:

CB23 1T score on the 258V
1859 (before the "bad" Windows update)
1819 (after the "bad" Windows Update)

Khato · Tuesday at 12:14 PM

GTracing said:
Overall good showing from Lunar Lake. Lunar lake pulls ahead of Meteor lake at higher graphics settings, but at lower settings it seems to have regressed in perf/watt in a couple games. I'm guessing the drivers need work.

Performance at lower settings is primarily a function of the lower power consumption limiting the CPU performance side. Could potentially be improved by modifying the power allocation between CPU and GPU, but maybe not meaningfully.

ikjadoon · Tuesday at 12:21 PM

ASUS Zenbook S 14 OLEDs power modes confirm, again, power profiles allow OEMs to exceed Intel's PL1.

Asus Zenbook S 14 OLED review (UX5406SA 2024 model, Intel Lunar Lake)

This is our review of the Asus Zenbook S 14 OLED UX5406S series, the mid-2024 addition to the Zenbook lineup of thin-and-light ultrabooks.

www.ultrabookreview.com

Whisper - 12W PL1
Balanced - 17W PL1
Performance - 24W PL1
Full Speed - 28W PL1

//

CPU Package power:

~3W to ~10W while browsing & text editing
~3W while Netflix

//

However, their 258V GB6.2 1T score is quite low: performance mode 1T is 2422.

GTracing · Tuesday at 12:23 PM

Khato said:
Performance at lower settings is primarily a function of the lower power consumption limiting the CPU performance side. Could potentially be improved by modifying the power allocation between CPU and GPU, but maybe not meaningfully.

I don't think that's the problem here. GTA 5 and X-Plane 11 are the only two games out of the 11 where this is happening, and X-Plane 11 is the only game where Lunar Lake loses to Meteor Lake at high or ultra settings.

CouncilorIrissa · Tuesday at 12:48 PM

So it looks like Lion Cove in LNL is the exact match for Zen 5 in STX.

Huang's article also contains the same benchmark @ 3GHz, but the image is too large to be attached.

FlameTail · Tuesday at 12:50 PM

poke01 said:
I was right. 15 watts @ 5.1 GHz

That is truly remarkable. This is as good as Snapdragon X Elite, if not better?

gdansk · Tuesday at 12:50 PM

CouncilorIrissa said:
So it looks like Lion Cove in LNL is the exact match for Zen 5 in STX.

Which should be enough. But I'm a bit worried - how many laptop makers will actually ship these with sensible power limits?

SiliconFly · Tuesday at 12:54 PM

StinkyPinky said:
These battery life tests with different displays and different display sizes etc just seem pointless. A 16" laptop with a brighter display drains quicker than a 14"? Who is surprised by that?

That said, I do like the GPU gains. Although once again, how many people are really gaming on thin and light laptops? Seems a very niche type of demand.

At this point, it's mostly about Lunar Lake's capability (than niche) where it beats competition in gaming even at significantly lower power.

Abwx said:
33% more cores and 25% higher power for a mere 11.8% better perf than a 2 years old CPU, quite a "win", if you set those 96C Epyc at the same 500W they would be on par, so if a 96C Zen 4 match Intel s 128C at same power i let you imagine what will happen soon with the Zen 5 based Epyc.

Abwx said:
Anyway based on these number we can infer that Zen 5 128C Epyc wil be largely ahead due to core count parity and also a slightly better process, add the augmented IPC and you get the picture

Yes, we can imagine and we get the picture. I'm waiting to see how much "massive" improvement Zen 5% is gonna bring!

I think the time when AMD parts being automatically superior is over. They now have to fight it out on all fronts considering they're falling behind in so many. Mismanagement? Lack of vision or leadership?

GTracing said:
... Overall good showing from Lunar Lake. Lunar lake pulls ahead of Meteor lake at higher graphics settings, but at lower settings it seems to have regressed in perf/watt in a couple games. I'm guessing the drivers need work.

Apparently, Battlemage drivers are far from final. They still need a lot of work.

naukkis said:
Very interesting results. Skymont low-power performance is unseen before on x86. Even @3.8GHz there ain't power budget for 4 P-cores. P-cores are doomed, intel should have put only 2 p-cores and add 2 or 4 e-cores to the ring for better MT perf.

Yep. Either the P cores have to evolve quickly or they're dead. But Intel is allocating more engineering resources to other teams by nixing ARL refresh. The next Cove should be better.

CouncilorIrissa said:
So it looks like Lion Cove in LNL is the exact match for Zen 5 in STX. ...

Excellent!

poke01 said:
I was right. 15 watts @ 5.1 GHz

Wow! 😍

fastandfurious6 · Tuesday at 1:01 PM

Intel’s entire turnaround plan hinges on this one new chip family – Clearwater Forest pictured, Intel’s first 18A chip slated for high-volume manufacturing

Clearwater Forest is key to Intel’s future.

www.tomshardware.com

title basically

fastandfurious6 · Tuesday at 1:02 PM

Abwx said:
33% more cores and 25% higher power for a mere 11.8% better perf than a 2 years old CPU, quite a "win", if you set those 96C Epyc at the same 500W they would be on par, so if a 96C Zen 4 match Intel s 128C at same power i let you imagine what will happen soon with the Zen 5 based Epyc.

yes, Turin = KO for Intel serverspace

Hitman928 · Tuesday at 1:02 PM

511 said:
Granite rappids review out destroying amd genoa we are so back

Intel Xeon 6980P "Granite Rapids" Linux Benchmarks Review - Phoronix

www.phoronix.com

Looks really good, though I will point out that a significant number of those tests are on Intel software. Will be interesting to see how well AMD can punch back with Turin. Definitely the most competitive it's been in quite a while though.

Hitman928 · Tuesday at 1:07 PM

FlameTail said:
That is truly remarkable. This is as good as Snapdragon X Elite, if not better?

Not quite as good as X Elite, though that's just an estimate as it doesn't have power sensors that can be read by HWinfo and the like.

511 · Tuesday at 1:20 PM

Hitman928 said:
Looks really good, though I will point out that a significant number of those tests are on Intel software. Will be interesting to see how well AMD can punch back with Turin. Definitely the most competitive it's been in quite a while though.

Only openVino but OpenVino is what fugaku supercomputer uses which is an ARM CPU they ported openVINO to Arm CPU embree is multi architecture Intel's SW Stack is something amd is unable to match even when intel was in trouble AMD may have better HW but Intel's software support is superior in industry there is no double also the funniest slide

Hitman928 · Tuesday at 1:29 PM

511 said:
Only openVino but OpenVino is what fugaku supercomputer uses which is an ARM CPU they ported openVINO to Arm CPU embree is multi architecture Intel's SW Stack is something amd is unable to match even when intel was in trouble AMD may have better HW but Intel's software support is superior in industry there is no double also the funniest slide View attachment 108155

OpenVINO, OSPRay, Embree, OpenVKL, and Open Image Denoise are all Intel software.

511 · Tuesday at 1:35 PM

Hitman928 said:
OpenVINO, OSPRay, Embree, OpenVKL, and Open Image Denoise are all Intel software.

They are not vendor locked to Intel and we use them in Blender/V ray corona and we know how they behave they are all open source I don't see the point outside of the use of AMX even for rigging purpose it's not CUDA 🤣

Hitman928 · Tuesday at 1:39 PM

511 said:
They are not vendor locked to Intel and we use them in Blender/V ray corona and we know how they behave they are all open source I don't see the point outside of the use of AMX even for rigging purpose

Of course they are not vendor locked, otherwise we wouldn’t have results on non-Intel CPUs. The point is that in the real world, Intel software doesn’t make up such a large portion of used software and of course Intel is going to optimize their own software for their own products. Just because it is open source, doesn’t mean equal optimizations for all architectures, especially with the “stock” software.

hemedans · Tuesday at 1:51 PM

CouncilorIrissa said:
So it looks like Lion Cove in LNL is the exact match for Zen 5 in STX.
View attachment 108154

Huang's article also contains the same benchmark @ 3GHz, but the image is too large to be attached.

At 4.6Ghz or below, From 4.6 and above LNC is 2-3% more efficient According to Huang.

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

Senior member

Attachments

Lifer

Senior member

Member

Member

Lifer

Senior member

Senior member

Senior member

Senior member

Golden Member

Senior member

Member

Senior member

Diamond Member

Platinum Member

Golden Member

Member

Member

Diamond Member

Diamond Member

Senior member

Diamond Member

Senior member

Diamond Member

Senior member