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.)

 

[Markfw]

I for one am very interested. It sounds like MAYBE Zen 5 will get a run for its money, I can hardly wait for benchmarks from BOTH as I don't trust marketing info from either company.

By Christmas I should know for sure. I assure you as a current Zen owner, I will be very honest when these come out and the benchmarks are available publicly.

Edit : this is the first time I have seen a chance for Intel to redeem themselves in many years. But both sides have convincing PR.

 

[Hulk]

I think we are going to see a lot of Lunar mobile devices Q3. I've seen a few demos of Lunar actually doing things for minutes at a time, running games and such, we as we know are crash prone, etc.. It makes me believe the silicon is solid. Plus TMSC has a good track record so the fab issues are kind of not so much there for this one.

Samples going to vendors now.. vendors figuring out new form factors for this tiny SoC, what they can do with the lower power envelope, higher performance, ... I think we might be (finally) pleasantly surprised with this one.

Once Intel guy in an interview said that Luner Lake with 8 threads is more performant than Meteor Lake with 22 threads. Not sure how that math works there but that is impressive.

 

[Henry swagger]

If Skymont IPC and overall performance in the too good to be true by this Fall really is true, let Intel Atom Austin team have all their Skymont cores in Arrow Lake SKUs and Lion Cove be scrapped.

Lion Cove having only 10-15% IPC over Sjkymont and only 600-800MHz clock speed advantage at triple or more the die size area, Israel Design Center does not deserve their cores in such case in the upcoming Arrow Lake SKUs.

But once again if the Skymont rumors of Raptor Cove IPC and maybe high 4.GHz clocks or even 5GHz all core feel too good to be true as fast as Fall 2024, it probably is.

But if what's too good to be true becomes true by this Fall. Intel please Skymnont only Arrow Lake SKUs with 16 or even more cores this fall on shelves and homogenous arch.

Who cares about only a small single thread performance uplift when you can get Golden Cove or Raptor Cove IPC or better with Skymont core 4.XGHz with X being high number or even 5GHz clocks if the too good to be true becomes amazingly true.

No really need for Lion Cove in above case especially since majority of market is mobile anyways. Just take it Intel losing single thread to AMD but a little bit especially if the Austin Atom team can evolve it even more to deliver Conroe knockout in a year or 2.

But once again if it feels too good to be true by this Fall it probably is. But maybe not???

A ex intel engineer said intel is loyal to idc even if atom surpass the p core in ipc..

 

[trivik12]

I think we are going to see a lot of Lunar mobile devices Q3. I've seen a few demos of Lunar actually doing things for minutes at a time, running games and such, we as we know are crash prone, etc.. It makes me believe the silicon is solid. Plus TMSC has a good track record so the fab issues are kind of not so much there for this one.

Samples going to vendors now.. vendors figuring out new form factors for this tiny SoC, what they can do with the lower power envelope, higher performance, ... I think we might be (finally) pleasantly surprised with this one.

Once Intel guy in an interview said that Luner Lake with 8 threads is more performant than Meteor Lake with 22 threads. Not sure how that math works there but that is impressive.

I think that is bcos chadmont > redwood cove. So its like 8 P cores

 

[Hulk]

I think that is bcos chadmont > redwood cove. So its like 8 P cores

Lunar Lake vs. Meteor Lake Compute
I can see 4 Lion Cove being equal to 6 Raptor Cove, best case for LNL assuming clocks are better for LNL, but it's a stretch.
I can also see 4 Skymonts being equal to 8 Crestmonts.

But then you have to make up for 6 logical cores and 2 LPE cores compute-wise?

 

[invisible_city]

A ex intel engineer said intel is loyal to idc even if atom surpass the p core in ipc..

I mean 1 random former employee is not a lot to go on. Leadership does change its mind. I think Intel is keen to fix its internal problems.

 

[Doug S]

I mean 1 random former employee is not a lot to go on. Leadership does change its mind. I think Intel is keen to fix its internal problems.

Yep, they have to be willing to toss out what doesn't work and double down on what does.

Like they did when they tossed out high power big core P4 design and went all in on what was intended as their low power laptop design that came to become known as "Intel Core". If their low power core is doing as well as it appears and their "P" core is stuck in the mud, maybe it is time to rethink where the investment dollars are going.

 

[invisible_city]

Yep, they have to be willing to toss out what doesn't work and double down on what does.

Like they did when they tossed out high power big core P4 design and went all in on what was intended as their low power laptop design that came to become known as "Intel Core". If their low power core is doing as well as it appears and their "P" core is stuck in the mud, maybe it is time to rethink where the investment dollars are going.

Yep

 

[SpudLobby]

Yep, they have to be willing to toss out what doesn't work and double down on what does.

Like they did when they tossed out high power big core P4 design and went all in on what was intended as their low power laptop design that came to become known as "Intel Core". If their low power core is doing as well as it appears and their "P" core is stuck in the mud, maybe it is time to rethink where the investment dollars are going.

Yes. PTL and beyond have a much more sane tile design, with fewer tiles. PTL has 3 tiles (CPU + SOC, GPU and PCD) + base Foveros 3D.

There should be a significant difference in pricing between Lunar Lake and Meteor Lake designs. So it should be fine.

Meteor Lake also scales up to H Gaming/Workstation Laptops, while Lunar is solely focused on premium ultrathin.

But really, Intel strategy is to have multiple overlapping product families in the market by 2025. Raptor, Meteor, Arrow, Lunar and Panther will all co-exist in the Mobile space next year.

Then, by 2026, Panther Lake should replace Meteor, Arrow and Lunar (Not really) and be the only product family alongside Raptor.

(There's also Wildcat Lake. But I dunno what market this thing is for)

So lunar as it’s won deal is probably kind of a one-off, I mean Panther Lake will replace it but you’ll also see a step back on the tiling (just nothing like MTL mess) and I assume on-package memory might be rolled back or something?

I can’t see Intel actually doing a Lunar Lake successor and some PTL U skus, so this is fine, as long as PTL has good enough battery and all.

 

[TESKATLIPOKA]

If Skymont IPC and overall performance in the too good to be true by this Fall really is true, let Intel Atom Austin team have all their Skymont cores in Arrow Lake SKUs and Lion Cove be scrapped.

Lion Cove having only 10-15% IPC over Sjkymont and only 600-800MHz clock speed advantage at triple or more the die size area, Israel Design Center does not deserve their cores in such case in the upcoming Arrow Lake SKUs.

But once again if the Skymont rumors of Raptor Cove IPC and maybe high 4.GHz clocks or even 5GHz all core feel too good to be true as fast as Fall 2024, it probably is.

But if what's too good to be true becomes true by this Fall. Intel please Skymnont only Arrow Lake SKUs with 16 or even more cores this fall on shelves and homogenous arch.

Who cares about only a small single thread performance uplift when you can get Golden Cove or Raptor Cove IPC or better with Skymont core 4.XGHz with X being high number or even 5GHz clocks if the too good to be true becomes amazingly true.

No really need for Lion Cove in above case especially since majority of market is mobile anyways. Just take it Intel losing single thread to AMD but a little bit especially if the Austin Atom team can evolve it even more to deliver Conroe knockout in a year or 2.

But once again if it feels too good to be true by this Fall it probably is. But maybe not???

What small ST uplift? Is 50% a small uplift?
I think Lion Cove should manage 5.8GHz and even If not, Skymont will not clock as high as 5GHz, you can forget about that.
According to Intel, Lion Cove supposedly provides >50% more performance in who knows what, that would mean <4.5GHz boost for Skymont in case of 5.8GHz for Lion Cove and 15% difference in IPC.

Even LNL has 4 Lion Coves and is a 15W CPU configurable to 30W, so 8 Lion Cove cores in a CPU with >55-125W TDP doesn't look like a problem. If you compared performance of 8 Lion cores vs 8 Skymont cores and had >=75W TDP(>=9W per core), then you would see a big difference in performance in my opinion.

Yeah, we can argue that spamming E-cores within the same area and power would provide better CB score, but even 8P+16E already has 24 cores and 32 threads including HT. How many of us can use this amount of threads? What SoC you want (>32 cores), looks more likely to happen for server than for desktop(laptop).

What I am interested in is, If Intel will do a Pentium with 8 Skymont cores or not? That would be a very capable SoC in my opinion.

 

[MarkizSchnitzel]

Which price segments is LNL targeting? Premium, or "premium midrange", if Arrow Lake will be Snapdragon X and Strix Point competitor?

 

[mikk]

So lunar as it’s won deal is probably kind of a one-off, I mean Panther Lake will replace it but you’ll also see a step back on the tiling (just nothing like MTL mess) and I assume on-package memory might be rolled back or something?

I can’t see Intel actually doing a Lunar Lake successor and some PTL U skus, so this is fine, as long as PTL has good enough battery and all.

In the leaked Dell roadmap next after Lunar Lake comes Nova Lake. Maybe Intel works on a real successor with this generation. PTL comes for the classic UPH segment.

 

[coercitiv]

Which price segments is LNL targeting? Premium, or "premium midrange", if Arrow Lake will be Snapdragon X and Strix Point competitor?

You can tell from the size and the node used that it has to aim for premium devices: the main die is 140mm2 on TSMC N3B, then adding to the cost is the second small die and the high-end packaging. There's no vertical integration to absorb some of the manufacturing cost, and the product needs to make money. Meteor Lake and Arrow lake refresh should take care of the lower segments.

 

[lightisgood]

In the leaked Dell roadmap next after Lunar Lake comes Nova Lake. Maybe Intel works on a real successor with this generation. PTL comes for the classic UPH segment.

Thinking realistically, there isn't clearly boundary between PTL (ARL) -U and LNL.
Based on Pat's saying, Intel only increased the parallelism of design teams in recent years.

 

[DavidC1]

Meteor Lake was basically a joke announcement so that Intel wasn't sued by investors. Designs only truly ramped by CES 24 and there isn't many MTL designs in the market still (when compared to a real Intel Mobile launch).

No. They needed Meteorlake to pipeclean Intel 4 and also the tile design.

@TESKATLIPOKA The leak said 5.8GHz for the P and 4.5GHz for the E did it not?

A ex intel engineer said intel is loyal to idc even if atom surpass the p core in ipc.

It's Gelsinger who is ultimately responsible for all of it. They did change things quite a bit better for the latter after he took over, even under financial stress. It looks like after Computex reveal maybe he'll live up to his former reputation.

 

[Mahboi]

Fortunately, there's a better slide comparing Skymont to Raptor Cove.

If that chart is real, it will be a good question about which core should be called P

Ok regarding these slides too:
- no numbers
- no numbered curve inversion point
What's with all the Apple-ness here? Just tell us at what watts do the P cores start being better. Which will also say when the E cores reach their viable max.
There's a lot of "just trust us breh" on what Intel's been showing. When do their first Skymont things come out? Cause I'd like to see those reviews...

 

[DavidC1]

There's a lot of "just trust us breh" on what Intel's been showing. When do their first Skymont things come out? Cause I'd like to see those reviews...

It's not a release so they're keeping it ambiguous. There's no real reason for them to be specific, even if they were people would be saying they are lying anyway.

Just wait for launch and reviews.

 

[FlameTail]

Here's the leaked Dell roadmap fwiw:


Lunar Lake in September. Panther Lake at end of 2025/early 2026.

 

[mikk]

So funny that they keep ignoring ARL mobile, it speaks for itself.

Qualcomm supports ARM's claim to have 50 % of the PC market in five years. Intel, of course, sees things differently.

"This is not the first Windows on ARM announcement. Lunar Lake and Panther Lake are very compelling products, so why would you not choose those."

https://twitter.com/x/status/1797858854950273311

 

[Ghostsonplanets]

Here's the leaked Dell roadmap fwiw:

View attachment 100729
Panther Lake at end of 2025?

Should be similar to Lunar Lake. Announcement at Computex, launch at Innovation Days and true ramp up with volume at CES 26.

 

[Ghostsonplanets]

So funny that they keep ignoring ARL mobile, it speaks for itself.

https://twitter.com/x/status/1797858854950273311

Arrow Lake is kinda of a filler release on Mobile, so it make sense.

 

[Mahboi]

It's not a release so they're keeping it ambiguous. There's no real reason for them to be specific, even if they were people would be saying they are lying anyway.

Just wait for launch and reviews.

Alright, it's September according to FlameTail.
That's pretty far, so I'm a bit disappointed. Those lofty promises won't be confirmed until then.

 

[Ghostsonplanets]

Updated Intel Meteor Lake Tuning For Linux Shows Huge Performance/Power Improvements - Phoronix

www.phoronix.com

Meteor Lake fine wine😅

 

[lightisgood]

Arrow Lake is kinda of a filler release on Mobile, so it make sense.

I agree with you.
LNL programs are working well now. So, ARL programs are prioritized for desktop.
This is IDM 2.0.

In addition, Intel 20A (ARL-H) is a pipe-cleaner for Intel 18A.
We saw a similar movement on Intel 4 / Intel3.
Intel didn't allocate large amount of wafer for MTL but for SRF&GNR.

 

[Hulk]

Does Arrow Lake mobile have the crappy 2-core LPE island on the SoC tile that’s fudgy and makes scheduling weirder? I wouldn’t want that lol.

I have read they’re ditching it but for desktop — true of mobile too, right? Probably they can afford to ditch it with better E Cores and maybe some Ring changes + N3.

In one of those videos with the engineers the Intel guy kind of admitted that while the point of the LPE Island was to be able to shut down most of the SoC during low compute workloads, but the LPE was to weak to really be useful. The much stronger (he said 4x) Skymont Island will be much more effective in this capacity for mobile. Meteor Lake was a test bed of sorts as I speculated earlier in this thread. They definitely learned from it for LNL.