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.)

SiliconFly · 2025-02-04T16:11:31-0500

dullard said:
Last rumor that I saw was more LP-E cores (4 instead of 2). And all cores get upgrades. Cougar Cove P cores instead of Lion Cove. E cores get a few optimizations, but aren't much different. LP-E cores should be able to be ran much faster (and hopefully will then be functionally usable) when not on TSMC 6.

What evidence do you have that this is no CPU improvement over ARL-H?

He's just trolling the Intel thread as usual. Ignore.

dullard · 2025-02-04T16:12:53-0500

poke01 said:
So apart from iGPU, PTL is not exciting?
Hopefully NVL-H is better otherwise AMD gonna eat Intel for lunch with their mobile designs.

See my comment above. But also 7.7x more TOPS if NPU performance matters to you.

adroc_thurston · 2025-02-04T16:26:48-0500

SiliconFly said:
Keep on dreaming.

Don't have to. It's a known product.

poke01 said:
So apart from iGPU, PTL is not exciting?

Yeah it's kinda the point really.

poke01 said:
Hopefully NVL-H is better otherwise AMD gonna eat Intel for lunch with their mobile designs.

I guess.

dullard said:
Last rumor that I saw was more LP-E cores (4 instead of 2). And all cores get upgrades. Cougar Cove P cores instead of Lion Cove. E cores get a few optimizations, but aren't much different

All of that costs power and power is a thing Intel already struggles with, see above.

SiliconFly · 2025-02-04T16:31:37-0500

adroc_thurston said:
Don't have to. It's a known product.

Obviously you don't. PTL does not share much with ARL.

adroc_thurston said:
All of that costs power and power is a thing Intel already struggles with

Thats what PTL addresses exactly.

adroc_thurston said:
see above.

Read abt PTL.

adroc_thurston · 2025-02-04T16:34:08-0500

SiliconFly said:
PTL does not share much with ARL.

Sure does, the suckass part (cores) are wholly derived.

SiliconFly said:
Thats what PTL addresses exactly.

It does not rofl, it moves the important stuff (CPU) to a worse node.

dullard · 2025-02-04T16:41:08-0500

adroc_thurston said:
It does not rofl, it moves the important stuff (CPU) to a worse node.

Okay, so now, where is your proof that 18A is a worse node than TSMC N3B? Considering you didn't bother to post any proof of your other unsubstantiated claims, I'm not very hopeful that you have anything here either.

MS_AT · 2025-02-04T16:41:45-0500

gdansk said:
PTL has a totally different uarch?

It's the industry secret but Intel has licensed M3 core from Apple in return for discount on the fabs to manufacture M5 if the tariffs hit to hard and TSMC is no longer viable. They will modify it to use x64 frontend. But don't tell anyone.

SiliconFly · 2025-02-04T16:47:40-0500

It's just typical trolling. Ignore.

One thing I've noticed consistently is, it's always okay to trash talk about Intel in a Intel thread, but it's NEVER okay to trash talk about AMD in a AMD thread. You get banned. I hope they fix this.

Mod callouts will get you points every time. Take it to the designated forum if you have beef. Aloha, enjoy the vacation.

Mod DAPUNISHER

igor_kavinski · 2025-02-04T16:52:16-0500

SiliconFly said:
One thing I've noticed consistently is, it's always okay to trash talk about Intel in a Intel thread, but it's NEVER okay to trash talk about AMD in a AMD thread. You get banned. I hope they fix this.

Then Intel needs to do better. Calling their worse cores worse shouldn't be a crime.

adroc_thurston · 2025-02-04T16:52:49-0500

SiliconFly said:
The Zen suckass cores are equally derivative.

They very clearly stomp the best Intel has to offer. See above.

SiliconFly said:
Dream on. It moves the important stuff (CPU) to a bleeding-edge world class node. 18A.

I admire your optimism but it's not quite that rosy.

SiliconFly · 2025-02-04T16:54:02-0500

adroc_thurston said:
I admire your optimism but it's not quite that rosy.

I admire your pessimism. But you just don't know stuff.

okoroezenwa · 2025-02-04T17:49:06-0500

SiliconFly said:
One thing I've noticed consistently is, it's always okay to trash talk about Intel in an Intel thread, but it's NEVER okay to trash talk about AMD in a AMD thread. You get banned. I hope they fix this.

Yeah, I doubt a “fix” is incoming for this. Them “safe space” rules go out of the window outside AMD threads for sure.

Hitman928 · 2025-02-04T17:55:50-0500

okoroezenwa said:
Yeah, I doubt a “fix” is incoming for this. Them “safe space” rules go out of the window outside AMD threads for sure.

You need to report the posts for action to be taken.

SiliconFly · 2025-02-04T17:56:58-0500

igor_kavinski said:
Then Intel needs to do better. Calling their worse cores worse shouldn't be a crime.

It's time for some PTL recap.

Slightly upgraded Panther Lake P core. Expected to bring in mid-single digit IPC increase.
Slightly upgraded Darkmont E cores. Some users mentioned that the front-end/decoder is being upgraded. Expected to bring in mid-single digit IPC increase.
IMC is being moved to the CPU tile to reduce latency.
SoC+GPU tile is being replaced with a single tile.
CPU tile is 18A. World's most advanced node in 2025.
All new 4x Darkmont LPE cores for improved efficiency.
Xe3 Celestial should outperform Xe2 Battlemage.
Better memory/ring clocks, cache latencies, etc.
Better NPU (more AI TOPS).
Better Packaging (improved Foveros for better performance/efficiency/bandwidth?)

C. C. Wei (TSMC) mentioned a while ago that N2 will ramp up in 2H 2025. But unfortunately, there were no takers due to various reasons (rumors include cost, yield, etc). None of the major players like Apple, Qualcomm, AMD, Intel who prefer leading-edge nodes opted to use N2. It got pushed to 2026, making TSMC N2 a total disaster in 2025. TSMC messed up. Bad for TSMC, good for Intel, cos Intel 18A is on track as confirmed multiple times by Intel.

Many didn't believe when Pat said Process Leadership in 2025, but it is happening as we speak. And PTL is the first product based on 18A to reach the market. PTL is expected to have more efficiency clock-for-clock compared to previous gen due to node jump and improved architecture. The math is simple. Thats one of the reasons Holthaus keeps saying PTL is similar to LNL rather than ARL.

Note: Also, 18A is the sole 2nm class node with real products in 2025. So, we can't compare 18A with N2 anymore, cos N2 is now 2026. And also Intel 18A-P is 2026. And among those two (18A-P & N2), 18A-P is the clear winner. Both are expected to have similar densities, but 18A-P is expected to have more performance, once again beating TSMC in process leadership in 2026.

Nova Lake is expected to use 18A-P in 2026. Some rumors say 14A, but I think it's a bit unlikely (but can't rule it out yet until this April). This combined with Coyote Cove & Arctic Wolf cores should give Intel an solid advantage considering Exist50 just mentioned recently that Nova Lake-SK is coming out with 48 cores (16P+32E). Nova Lake should have good efficiency and terrific performance.

For comparison, Zen 6 client is on a N3 variant I think (and not N2). One generation behind.

DAPUNISHER · 2025-02-04T18:05:50-0500

okoroezenwa said:
Yeah, I doubt a “fix” is incoming for this. Them “safe space” rules go out of the window outside AMD threads for sure.

The only place to discuss moderation is the designated forum for it.

And you might want to put down the haterade and take a better look. Offender has been vacationed and a public warning issued. But don't let that get in the way of your false narrative about how this is an AMD biased forum.

FYI: Along with questioning moderation being a rule violation, so is responding to official mod posts (you can tell because they are signed and bolded) so don't do it.

Mod DAPUNISHER

Hitman928 · 2025-02-04T18:14:26-0500

SiliconFly said:
It's time for some PTL recap.

Slightly upgraded Panther Lake P core. Expected to bring in mid-single digit IPC increase.

Slightly upgraded Darkmont E cores. Some users mentioned that the front-end/decoder is being upgraded. Expected to bring in mid-single digit IPC increase.

IMC is being moved to the CPU tile to reduce latency.

SoC+GPU tile is being replaced with a single tile.

CPU tile is 18A. World's most advanced node in 2025.

All new 4x Darkmont LPE cores for improved efficiency.

Xe3 Celestial should outperform Xe2 Battlemage.

Better memory/ring clocks, cache latencies, etc.

Better NPU (more AI TOPS).

Better Packaging (improved Foveros for better performance/efficiency/bandwidth?)

C. C. Wei (TSMC) mentioned a while ago that N2 will ramp up in 2H 2025. But unfortunately, there were no takers due to various reasons (rumors include cost, yield, etc). None of the major players like Apple, Qualcomm, AMD, Intel who prefer leading-edge nodes opted to use N2. It got pushed to 2026, making TSMC N2 a total disaster in 2025. TSMC messed up. Bad for TSMC, good for Intel, cos Intel 18A is on track as confirmed multiple times by Intel.

Many didn't believe when Pat said Process Leadership in 2025, but it is happening as we speak. And PTL is the first product based on 18A to reach the market. PTL is expected to have more efficiency clock-for-clock compared to previous gen due to node jump and improved architecture. The math is simple. Thats one of the reasons Holthaus keeps saying PTL is similar to LNL rather than ARL.

Note: Also, 18A is the sole 2nm class node with real products in 2025. So, we can't compare 18A with N2 anymore, cos N2 is now 2026. And also Intel 18A-P is 2026. And among those two (18A-P & N2), 18A-P is the clear winner. Both are expected to have similar densities, but 18A-P is expected to have more performance, once again beating TSMC in process leadership in 2026.

Nova Lake is expected to use 18A-P in 2026. Some rumors say 14A, but I think it's a bit unlikely (but can't rule it out yet until this April). This combined with Coyote Cove & Arctic Wolf cores should give Intel an solid advantage considering Exist50 just mentioned recently that Nova Lake-SK is coming out with 48 cores (16P+32E). Nova Lake should have good efficiency and terrific performance.

For comparison, Zen 6 client is on a N3 variant I think (and not N2). One generation behind.

AFAIK, N2 is and always was scheduled to enter HVM at the end of this year.

OneEng2 · 2025-02-04T21:28:31-0500

511 said:
View attachment 116268
DMR Launching Q2/Q3 26?

Possibly, but not sure how H2 2026 is Q2? Furthermore, recent trends at Intel have not indicated early completion of their roadmap, but rather slippage in the timeline. At best, H2 2026 in current Intel speak (using the best empirical evidence we have from other recent announcements) at best means December 2026. Let's see how PTL does on 18A ... and then CWF before we start making early predictions on DMR.

dttprofessor said:
AMD's good days in mid-to-high-end notebooks are over, even with an 8-10% IPC limitation on Pcore@285H (compared to 258V), ULTRA2 still has considerable competitiveness. The only competitive products AMD currently has in the laptop market are the low-end 7845H and 8845H.

I just went over the AMD quarterly results. I thought their laptop share went up? Someone correct me if I am off base here.

dullard said:
Okay, so now, where is your proof that 18A is a worse node than TSMC N3B? Considering you didn't bother to post any proof of your other unsubstantiated claims, I'm not very hopeful that you have anything here either.

Since there is nothing built on 18A yet, there can't be any "proof". I would say that it is highly likely that 18A will be superior to N3B in nearly every aspect ..... except perhaps fully loaded cost to Intel. Hard to absorb that 20bn development price tag amortization and not end up with an insanely expensive butchers bill.

511 · 2025-02-04T21:49:43-0500

OneEng2 said:
Possibly, but not sure how H2 2026 is Q2? Furthermore, recent trends at Intel have not indicated early completion of their roadmap, but rather slippage in the timeline. At best, H2 2026 in current Intel speak (using the best empirical evidence we have from other recent announcements) at best means December 2026. Let's see how PTL does on 18A ... and then CWF before we start making early predictions on DMR.

I just went over the AMD quarterly results. I thought their laptop share went up? Someone correct me if I am off base here.

Since there is nothing built on 18A yet, there can't be any "proof". I would say that it is highly likely that 18A will be superior to N3B in nearly every aspect ..... except perhaps fully loaded cost to Intel. Hard to absorb that 20bn development price tag amortization and not end up with an insanely expensive butchers bill.

Well the only thing N3B has over 18A is SRAM Density not PPA

jdubs03 · 2025-02-04T23:36:46-0500

Hitman928 said:
AFAIK, N2 is and always was scheduled to enter HVM at the end of this year.

Let’s just say Panther Lake will be on 18A releasing roughly in late October - early November.

Then that would indeed be a shipping product on a leading edge node. Since N2 wouldn’t be out.

So then what is the first product to release on N2?
I was thinking maybe the Snapdragon 8 Elite 2 but that might be too early. It’s not the A19(Pro), doubtful it’s the M5 variants since Apple usually has the A and M series on the same process.

Maybe a Mediatek 9500+ in spring 2026?

Tbh it might be worth collaborating on a timeline for node/product release. Could clear up some ambiguity.

Hitman928 · 2025-02-04T23:43:21-0500

jdubs03 said:
Let’s just say Panther Lake will be on 18A releasing roughly in late October - early November.

Then that would indeed be a shipping product on a leading edge node. Since N2 wouldn’t be out.

So then what is the first product to release on N2?
I was thinking maybe the Snapdragon 8 Elite 2 but that might be too early. It’s not the A19(Pro), doubtful it’s the M5 variants since Apple usually has the A and M series on the same process.

Maybe a Mediatek 9500+ in spring 2026?

Tbh it might be worth collaborating on a timeline for node/product release. Could clear up some ambiguity.

Spring would probably be too quick, though it depends on just how late in the year it goes HVM. Late summer to fall time is more likely.

Khato · 2025-02-05T00:32:53-0500

OneEng2 said:
Since there is nothing built on 18A yet, there can't be any "proof". I would say that it is highly likely that 18A will be superior to N3B in nearly every aspect .....

Yup, there is no 'proof' in the public domain as of yet as all products built on 18A thus far are unreleased. The only points which potentially stand in the way of 18A are yield and available capacity.

igor_kavinski · 2025-02-05T00:55:16-0500

Intel 3 vs. Intel 4 looks decent though there's a much improved architectural overhaul involved: https://browser.geekbench.com/v6/cpu/compare/9830182?baseline=10097869

Those multicore regressions are hopefully just early BIOS issues.

DavidC1 · 2025-02-05T01:57:04-0500

igor_kavinski said:
Intel 3 vs. Intel 4 looks decent though there's a much improved architectural overhaul involved: https://browser.geekbench.com/v6/cpu/compare/9830182?baseline=10097869

Those multicore regressions are hopefully just early BIOS issues.

There's no architectural improvement, it's nearly entirely due to process. Core Ultra 7 265U is Meteorlake cores on Intel 3. The ST clocks are 10% higher so that's why the gains are close to it. As for MT, Geekbench does vary a ton(it's a userbenchmark after all), so we'll have to compare top result to top result. It should be 10% or more faster.

DavidC1 · 2025-02-05T02:08:46-0500

511 said:
Well I don't think so cause they have not benefited from AI as much other companies

Sure they will. All computers are benefitting from it.

Also a quote from David Huang:

If both the large core and the small core of 255H and HX 370 are set at 3.8 GHz, it can be found that the large core of this generation is only less than 8% ahead of the same frequency performance of the Skymont small core.

Skymont in Arrowlake is 34% faster in Int per clock compared to Crestmont in Meteorlake.

SiliconFly said:
(16P+32E). Nova Lake should have good efficiency and terrific performance.

Intel for the past 10 years has should have rebranded themselves with the motto "only rosy years are in the future".

511 · 2025-02-05T02:59:36-0500

DavidC1 said:
Sure they will. All computers are benefitting from it.

they will benifit from Win10 eol than AI

DavidC1 said:
Also a quote from David Huang:

Skymont in Arrowlake is 34% faster in Int per clock compared to Crestmont in Meteorlake.

It's hillarious that a tiny core is so much capable

DavidC1 said:
Intel for the past 10 years has should have rebranded themselves with the motto "only rosy years are in the future".

Well since the node debacle the focus was always their node their design issues are only showing for the past two years after the nodes were fine.

I am still in awe that design teams were allowed near Infinte tape outs and it was the foundry's that has to accommodate the changes not design.

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

Senior member

Attachments

Golden Member

Elite Member

Diamond Member

Golden Member

Diamond Member

Elite Member

Senior member

Golden Member

Lifer

Diamond Member

Golden Member

Member

Diamond Member

Golden Member

Super Moderator CPU Forum Mod and Elite Member

Diamond Member

Senior member

Golden Member

Golden Member

Diamond Member

Golden Member

Lifer

Golden Member

Golden Member

Golden Member