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	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.)

lightisgood · Jun 6, 2024

Hulk said:
Is the memory side cache the 8MB "mystery cache?"

Please remember Skylake, eDRAM.
In some workloads, Skymont may be given priority memory-side$ allocation.
That's all.

The Intel Skylake Mobile and Desktop Launch, with Architecture Analysis

www.anandtech.com

deasd · Jun 6, 2024

Joe NYC said:
I have a hunch hat it will go the other direction. Taking just one variable, memory latency and some round numbers:

MTL P-Core to LNL P-Core = +14%
(given from the presentation)

This transition improves latency since memory controller is on the same die in LNL (and also SLC). Let's say, memory latency improvement buys 4%, then excluding the memory latency improvement, core for core is 10%.

This, is what I thought.

1.LNL use LPDDR5X-8533
2.memory module is side by side with CPU package which means low latency, similar design to Apple M series
3.memory controller is integrated on die

These gives LNL huge advantages on memory sensitive tests, especially some from SPEC
You could buy faster memory for DT platform, but 2 and 3 advantage are the other Arrowlake platform couldn't have.

TESKATLIPOKA · Jun 6, 2024

SiliconFly said:
P cores still the main cores. E cores kick in only when the s/w threads don't require that much performance and/or idling a bit. E cores are still "crutches" or just "helper" cores to the main P cores, cos ST perf is still the most important factor when it comes to performance.

Even in basic tasks like web browsing, ST perf is still very crucial.

Lion Cove does have higher ST perf than Skymont, but the question is at what power, don't forget LNL has only 17W TDP configurable to 30W.

For 1-2 threaded APPs P-core will be faster, but what about more?
Can a Lion Cove cluster outperform a Skymont cluster at 17W TDP or does It need 30W?

Basically my question is at what power does this P-core end up more efficient than E-core.

coercitiv · Jun 6, 2024

TESKATLIPOKA said:
Lion Cove does have higher ST perf than Skymont, but the question is at what power, don't forget LNL has only 17W TDP configurable to 30W.

Hulk · Jun 6, 2024

coercitiv said:
View attachment 100651

Gotta love how Intel can create a microprocessor with billions of transistors but they have such a hard time putting numbers and units on the axis of graphs.

The graphs were devised by cavemen, "This direction bigger, this one smaller."

Yes, yes, I know they intentionally obscure actual numbers and are only looking to show trends. It's just frustrating.

deasd · Jun 6, 2024

About supply:

Lee Yoon-seok, head of LG Electronics' IT division, visited Computex and said, "I am worried about the delay in supplying Intel Luna Lake"

deepl

‘We are worried that Intel is about half a year behind AMD and Qualcomm (Snapdragon X Elite),’ Lee said of Intel's schedule for launching new products with its laptop processors, which will be available from March.

google

Regarding the release schedule for new products equipped with the laptop processor 'Luna Lake' that Intel will supply starting in March, business manager Lee Yoon-seok said, "We are concerned that Luna Lake is about half a year behind AMD and Qualcomm (Snapdragon X Elite)."

https://twitter.com/x/status/1798649833894609134

ToTTenTranz · Jun 6, 2024

deasd said:
About supply:

https://twitter.com/x/status/1798649833894609134

Wasn't the Claw 8 announced to be shipping in September? Or was that just people making assumptions?

Regardless, it's a bit weird that they were showing working units to the Computex press in early June if it's only going to launch 9 months later.

deasd · Jun 6, 2024

ToTTenTranz said:
Wasn't the Claw 8 announced to be shipping in September? Or was that just people making assumptions?

Regardless, it's a bit weird that they were showing working units to the Computex press in early June if it's only going to launch 9 months later.

Google is added. It seems Deepl translate something wrongly. I believe Google is correct one.

invisible_city · Jun 6, 2024

If Intel is really launching lunar lake in 9 months that would be deeply worrying given the fact that they’re saying that it is shipping to OEMs now.

Hitman928 · Jun 6, 2024

deasd said:
About supply:

deepl

google

https://twitter.com/x/status/1798649833894609134

With QC and AMD laptops launching within the next 6 weeks or so (QC this month), they'll make the back to school season and obviously holiday season. With Lunar Lake launching later, they'll for sure miss the back to school season and maybe the gentleman from LG is worried that they won't be able to get Lunar Lake laptops out the door even in time for the holiday rush, at least not in good volume. That's my interpretation of it anyway.

Hitman928 · Jun 6, 2024

invisible_city said:
If Intel is really launching lunar lake in 9 months that would be deeply worrying given the fact that they’re saying that it is shipping to OEMs now.

Depends on what they're shipping. If they are just shipping final QS, then the schedule might be tight for getting the final products out in volume before the holiday season.

eek2121 · Jun 6, 2024

inf64 said:
Based on what exactly (20%)? Is that your hopium/copium or you base that on some benchmarks? We already have a chart from intel that shows Lunar Lake vs RWC (which has lower IPC than RPC) and the chart is ~14% average, with even a regression in some benchmarks.

From AMD, we have a very vague average with very conservative workload selection that shows 16% improvement. I'm very confident that Zen 5 will have no problem beating Arrow Lake in ST stuff and gaming, while MT stuff remains open due to E cores and huge uplift they bring.

The P-Core i. Arrow Lake will be different from the one in Lunar Lake. It will have more cache and some other changes.

ondma said:
Dont forget there will be a clock speed regression of 5-10% at least, and a IPC 20% seems a bit high. So final result, I would think 10% faster than RC ST would be an acceptable result. As for MT, if the E cores are as good as the predictions, there should be plenty of horsepower for a good MT performance. It will depend on temps and power usage, i.e. what clocks they can hold in full core load.

No single core clock regression has been noted so far.

H433x0n said:
Depends on if the leaked figures for ARL-S QS silicon is correct, those figures would suggest a 5% clock regression.

As of now the 8+32 SKU is dead.

8+32 would be a huge waste of die area without increasing the power limit. Arrow Lake has a reduced power limit compared to Raptor Lake. Not happening. Instead, Intel is focusing on the E-core IPC.

Hulk said:
I just noticed that the slide showing Skymont outperforming Raptor Cove by 2% in "general software" there is a little note reading "when on the LLC ring."

So Skymonts in LL are not Raptor Coves. But they will be in ARL.

Yep, baseline E core performance is now +-10% where current P-Cores are.

Wolverine2349 · Jun 6, 2024

Yep, baseline E core performance is now +-10% where current P-Cores are.

Click to expand...

IPC or overall performance including clocks?

TESKATLIPOKA · Jun 6, 2024

coercitiv said:
View attachment 100651

I saw this graph, but It is useless without any numbers.
Do we know at what power do they have the same perf/W ratio? No

Philste · Jun 6, 2024

ToTTenTranz said:
Regardless, it's a bit weird that they were showing working units to the Computex press in early June if it's only going to launch 9 months later.

My guess is that this person confused it with Arrow Lake mobile, which should be a CES2025 launch. After all, Arrow Lake is the product that competes with Strix and X Elite, Lunar Lake is one Performance (and power) tier lower. I mean literally everyone said September for Lunar Lake even a few months ago.

Joe NYC · Jun 6, 2024

Philste said:
My guess is that this person confused it with Arrow Lake mobile, which should be a CES2025 launch. After all, Arrow Lake is the product that competes with Strix and X Elite, Lunar Lake is one Performance (and power) tier lower. I mean literally everyone said September for Lunar Lake even a few months ago.

Lunar Lake name is in the direct quote.

But this is only one source, one datapoint. Best to mentally just file it in its category with a single datapoint suggesting it, until there is another independent datapoint.

Hulk · Jun 6, 2024

TESKATLIPOKA said:
I saw this graph, but It is useless without any numbers.
Do we know at what power do they have the same perf/W ratio? No

I'm on my phone now so can't dig up the link. But earlier today I saw a video of a meteor lake in lunar lake laptop running a game both at 60 frames per second. The lunar lake power consumption for the soc was nearly half that of meteor lake. Perhaps 75% at times. Still this is enormous especially when you consider that this included memory power consumption for lunar lake but not for meteor lake.

lightisgood · Jun 6, 2024

deasd said:
About supply:

https://twitter.com/x/status/1798649833894609134

Intel's Dan Rogers said that...

1. LNL is in the QS stage.
2. Intel will launch LNL in 3Q24. Almost LNL programs are on schedule.
3. LNL is targeted for holiday season.

I don't know what he worry about LNL supply.

StinkyPinky · Jun 6, 2024

I mean it's kinda weird we know nothing about the actual CPU sku's. AMD and Qualcomm have released pretty much all the details for theirs.

Intel messed up again didn't they? I feel like a child whose father promised to take them to a ball game but instead got in a week long coke binge and ended up in rehab instead.

Intel, intel, intel.

Khato · Jun 6, 2024

Well, when you consider that Qualcomm's initial big splash for X-elite was in October of 2023 and they're supposed to finally be available in mid June of 2024... Yeah, it'd be hard for Intel to do worse.

That said, we probably are going to see a repeat of the MTL launch where there are a few 'early release' designs available for purchase on launch day in Q3. The vast majority of designs will probably be released 2-3 months later, somewhat in time for the holiday season.

trivik12 · Jun 6, 2024

https://twitter.com/x/status/1798699363763699760

Raichu is saying LNC is held back by LNL.

" it is almost not the problem and the comparison is in worse conditions and not full-configuration.(LNL has less L3+SLC, low power fabric, and 2.5M instead of 3M L3)"

Henry swagger · Jun 6, 2024

trivik12 said:
https://twitter.com/x/status/1798699363763699760

Raichu is saying LNC is held back by LNL.

" it is almost not the problem and the comparison is in worse conditions and not full-configuration.(LNL has less L3+SLC, low power fabric, and 2.5M instead of 3M L3)"

Desktop lion cove will have higher ipc.. mayb 5% more ?

trivik12 · Jun 6, 2024

16% more than Raptor Cove would be hilarious.

Henry swagger · Jun 6, 2024

trivik12 said:
16% more than Raptor Cove would be hilarious.

In what way ?

coercitiv · Jun 6, 2024

trivik12 said:
Raichu is saying LNC is held back by LNL.

In some ways it's holding it back, in others it's boosting it. Smaller cache and slower fabric is a penalty, but on-package memory is a boon. Memory controller being on the same tile is also a big plus.

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

Senior member

Attachments

Senior member

Senior member

Platinum Member

Diamond Member

Diamond Member

Senior member

Member

Senior member

Member

Diamond Member

Diamond Member

Diamond Member

Senior member

Platinum Member

Member

Platinum Member

Diamond Member

Senior member

Diamond Member

Attachments

Golden Member

Senior member

Senior member

Senior member

Senior member

Diamond Member