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.)

 

[Abwx]

Exactly. I think when Intel says 2X, I think it's under normal circumstances.

Intel themselves said 40% lower power at same frequency for Intel 4 vs Intel 7 at 2.1GHz, that s 1.67X at low frequency.

Above 3GHz that s more like 10-15% better perf according to AT, this amount to a 1.25x-1.4x range for power reduction at same frequency.

Laws of physics are what they are and are...

Still quoting the baboon? I'll repeat. MTL's so-called 2X power efficiency is due to its new superior hyper-efficient architecture and not the node jump alone.

Just fyi, do you know that MTL's NOC Fabric is far superior to AMD's Infinity fabric? It can deliver up to 4X the bandwidth under certain conditions compared to competition. Also very efficient. MTL's efficiency comes from lots of under-the-hood improvements. Think you missed Intel Innovation.

Architecture will do nothing about it, that still RPL uarch even if IPC improved, actually higher IPC will even reduce intrinsical efficency, no CPU is immune from this drawback including AMD ones, their latest uarch is less efficient than the previous one but the new process hide the thing.

As for MLID whatever his alleged sources his numbers are close to what can be expected from Intel 4.

 

[SiliconFly]

Intel themselves said 40% lower power at same frequency for Intel 4 vs Intel 7 at 2.1GHz, that s 1.67X at low frequency.

Above 3GHz that s more like 10-15% better perf according to AT, this amount to a 1.25x-1.4x range for power reduction at same frequency.

Laws of physics are what they are and are...



Architecture will do nothing about it, that still RPL uarch even if IPC improved, actually higher IPC will even reduce intrinsical efficency, no CPU is immune from this drawback including AMD ones, their latest uarch is less efficient than the previous one but the new process hide the thing.

As for MLID whatever his alleged sources his numbers are close to what can be expected from Intel 4.

Architecture has everything to do about it.

MTL has lots of power-saving tech under-the-hood. That 20% PPW gain from Intel 7 to Intel 4 applies only to the RWC tCPU tile. MTL is a SoC. Not just the tCPU tile alone.

Kindly watch Intel Innovation...

 

[Glo.]

MTL has lots of power-saving tech under-the-hood. That 20% PPW gain from Intel 7 to Intel 4 applies only to the RWC tCPU tile. MTL is a SoC. Not just the tCPU tile alone.

Kindly watch Intel Innovation...

How do you know that considering that it has not been released yet, and all we have is Intel marketing?

 

[SiliconFly]

How do you know that considering that it has not been released yet, and all we have is Intel marketing?

The only thing we know for sure at this point is what Intel has shared. Like the new power-optimized Intel 4 HP cells - tCPU only, LP E-cores, DLVR, NOC, powering-down independent tiles, an efficient SoC tile design that can run totally independent (eg, while watching youtube), etc. Intel has claimed on multiple occasions that MTL SoC (not tCPU alone) can deliver up to 2X efficiency (probably under some normal workloads). Rest all are rumors at this point based on unreliable leaks or speculations. Given all the data, it looks more like MTL is actually a lot more efficient that previous gen. Maybe 2X. Maybe not. But that only time can tell.

 

[Mopetar]

MLIDs video on the topic of Meteor Lake desktops and is leaking Cinebench score of MTL 155H SKU.

P.S. That power draw is awful.

Usual disclaimers about MLID apply, but I'd want to know more about the settings used for the run. Intel has had several generations where they'll let their chips go all out for an extra 5% even if the efficiency drops 30% or more to get there.

 

[Geddagod]

A 6 + 8 13700H use close to 89W to score 16000 and 51W to score about 13300 :

https://www.notebookcheck.net/Dell-Inspiron-16-Plus-7630-laptop-review-60-W-GeForce-RTX-4060-isn-t-so-bad.743747.0.html
So much for your 16k/ 65W, reality is 30-37% more power than your flawed assumptions.

No, I quoted quite literally from notebook check too lol. I even said so in my original response. It's not an "assumption". That power curve graph I have shown is from this notebookcheck review comparing the 7940hs vs the 13700h.

Here the 13700H page at NBC, you can see the powers and scores in CB R23, none is even remotely close of your statements :

Blud it's not even "my" statements. You think I made up the graph or something? Lol.

21.5% better perf at isopower is at 2.1GHz, as frequency increase the gain is shrinking, there s an article at AT

Eh I just used 25% bcuz IIRC that was the original claim I said would be enough, and you said it wasn't lol. Looking at the graph, at 35 watts Intel needs a 17% gain in perf/watt, 25% at 45 watts, 20% at 55 watts, and 10% at 65 watts. These are very rough estimations based on the power scaling graph.

As for cheap AMD Ryzen that just plain deffamation, the market is full of half baked Intel laptops, or rather furnaces, find me a single Intel laptop at this price that has as well balanced perfs and caracteristics :

Nah, don't think @SiliconFly is wrong abt that. Intel has way more options, models, and IIRC from when I last went laptop shopping, cheaper prices than AMD laptops too. I do recall one really good deal on a AMD laptop with a combo AMD CPU and GPU, but the Power limit for the GPU was pathetic (and it only came with a 512?GB SSD) so I didn't go for it. Quite simply, Intel dominates the OEM market vs AMD. Sure, AMD might be the better choice, but relatively few laptop models contain AMD CPUs bcuz Intel has OEMs by the balls.

A 6nm fabbed 6850U score 10 000 pts at 30W, a 13700H score 12 000 pts at 45W, so at 30W it would barely reach the 10k mark, actually even at 6 + 8 RPL mobile is somewhat less efficient than a 8C Rembrandt,

Rembrandt looks to have the same or worse perf/watt than ADL mobile across the entire power curve from 35 watts and up

Also seen here

what about if we pit a 6 + 4 instead, wich is the real comparison..?..

Omg. 6+8 is the real comparison for an 8 core CPU. If Intel had no E-cores, they would slot in 2 P-cores in the same exact heckin' spots they put their 4 e-core clusters. I have posted the die shots of these CPUs like a day ago, idk if you just chose to ignore them or what lol. Literally everyone use the 4 E-core = 1 P core model... because it's literally true, so idk why you insist on being wrong on this one point.

 

[Geddagod]

actually higher IPC will even reduce intrinsical efficency, no CPU is immune from this drawback including AMD ones, their latest uarch is less efficient than the previous one but the new process hide the thing.

Idk what you mean by "intrinsical efficiency" but Zen 4, iso node, was actually a slight uplift in perf/watt over Zen 3. Meaning even if they didn't move to N5 and uses the same node as Zen 3, the perf/watt would improve.
Zen 3 also improved frequency iso power vs Zen 2, despite its large IPC improvement.

 

[Geddagod]

As for future Intel products MLID leak state a 7300 score in CB R20 for a 6 + 8 MTL, allegedly at 100W.

7300 pts in R20 amount to about 18700-19000 pts in CB R23, a comparable RPL, that is a 6 + 8 13700H, does 16000 pts at about 89W, so that s about the 18% better perf/isowatt that is brought by Intel 4.

According to AT Intel s 4 21.5% better perf/isowatt is at 2.1GHz, as frequency increase the percentage shrink, so a 18% figure above 3GHz is realistic.

These numbers make 0 sense at all. MLID's numbers suggest MTL will bring no perf/watt gain at all. There's pretty much four options I see here:
1. Intel 4 is a failure
2. However MLIDs OEM source measured power consumption, is wrong
3. MLID is wrong
4. This Pre-QS is simply force fed power to ensure that the sample is hitting intended clocks
Could be any combination of these reasons, or none at all. But let's not pretend these results make sense

A 12700h btw scores 18k points at 95 watts (HWUB). Scoring 19k at 100 watts indicate a 5% uplift in perf/watt.

 

[Abwx]

A 12700h btw scores 18k points at 95 watts (HWUB). Scoring 19k at 100 watts indicate a 5% uplift in perf/watt.

That s wrong, he didnt do an accurate measurement for the 12700H, a 13700H, wich has notably better perf/watt, cant break the 18k unless set at 115W at least :

https://www.notebookcheck.net/Intel-Core-i7-13700H-Processor-Benchmarks-and-Specs.677405.0.html

 

[Geddagod]

That s wrong, he didnt do an accurate measurement for the 12700H,

Don't think so, if you look at the two graphs I've shown you, the 12700h and 13700h graphs from notebookcheck and HWUB nearly match eachother perfectly, which is expected

a 13700H, wich has notably better perf/watt

Pretty sure a 13700h is just a better binned 12700h.... don't think it's "notably better"

ant break the 18k unless set at 115W at least :

One of the laptops scores 17.4k points while having a upper TDP limit of 80 watts. Another one scores 18k+ from 110 watts. But we also have a ton of other sources showing a 13700h can score ~18k points without going too much above, or even above, 100 watts:



In reality, a score of ~7300 from MTL using 100 watts is not impressive at all, and indicates little to no perf/watt uplift over last gen at that power.
Looking at perf/watt, at a 100 watts, the 13700h/12900h is already scoring ~18k points. This would mean that MTL is a ~5% improvement in perf/watt at 100 watts.

 

[aigomorla]

@Abwx You are getting VERY close to inciting and flame baiting. I have to agree with other a few other posters that you are cherry picking and then trying to pass AMD propaganda, instead of technical. Do it again and I will exit you from future participation of this thread permanently.

@SiliconFly The only people that are allowed to tell someone to leave the thread is us moderators, do it again, and you will be given an infraction. Everyone is entitled to there opinion, don't like it, then put them on ignore. If we find they are distructive to chat, or inciting, we will force a exodus on them and ban them from future participation in the thread.

Moderator Aigo

 

[Abwx]

One of the laptops scores 17.4k points while having a upper TDP limit of 80 watts. Another one scores 18k+ from 110 watts. But we also have a ton of other sources showing a 13700h can score ~18k points without going too much above, or even above, 100 watts:

I guess that the laptop you re talking about is this one :

https://www.notebookcheck.net/Lenovo-Yoga-Pro-7-14-G8-notebook-review-GeForce-RTX-4050-laptop-powers-3K-display-with-120-Hz.715978.0.html

So let s look at the details of the numbers, they test with CB R15 and measure 80W or so with the software monitoring, but now let s look at the measurment they do on the main with CB R15 and external monitor where the laptop use 112W for CB R15, wich amount to 80W at the CPU level, so everything seems right isnt it..?..

But the same main measurement show 126W, 14W more than R15, for the test using CB R23, this amount to a difference of 10W at the CPU level, hence the CPU use roughly 90W in R23 and still doesnt break the 18k, to do so frequency must be increased by 3.4% wich will drive the CPU to 100W.

So ultimately MTL has 9% higher perf at isopower, and that s with an ES chip wich, as i already stated, is eventually overvolted for tests by at least 5%, wich will eat as much in the actual improvement, in other words the perfs is improved by about 15% at isopower, that s Intel 4 stated improvement in the 3-3.5GHz range, namely the one at wich this 13700H is running in MT....

Edit : According to AT perf improvement at isopower in this frequency range, 3 to 3.5, is 10% with standard cells and can be pushed to 15% with HP cells.

Intel 4 Process Node In Detail: 2x Density Scaling, 20% Improved Performance

www.anandtech.com

 

[H433x0n]

No, I quoted quite literally from notebook check too lol. I even said so in my original response. It's not an "assumption". That power curve graph I have shown is from this notebookcheck review comparing the 7940hs vs the 13700h.

Blud it's not even "my" statements. You think I made up the graph or something? Lol.

Eh I just used 25% bcuz IIRC that was the original claim I said would be enough, and you said it wasn't lol. Looking at the graph, at 35 watts Intel needs a 17% gain in perf/watt, 25% at 45 watts, 20% at 55 watts, and 10% at 65 watts. These are very rough estimations based on the power scaling graph.

Nah, don't think @SiliconFly is wrong abt that. Intel has way more options, models, and IIRC from when I last went laptop shopping, cheaper prices than AMD laptops too. I do recall one really good deal on a AMD laptop with a combo AMD CPU and GPU, but the Power limit for the GPU was pathetic (and it only came with a 512?GB SSD) so I didn't go for it. Quite simply, Intel dominates the OEM market vs AMD. Sure, AMD might be the better choice, but relatively few laptop models contain AMD CPUs bcuz Intel has OEMs by the balls.

Rembrandt looks to have the same or worse perf/watt than ADL mobile across the entire power curve from 35 watts and up
View attachment 86246
Also seen here
View attachment 86247

Omg. 6+8 is the real comparison for an 8 core CPU. If Intel had no E-cores, they would slot in 2 P-cores in the same exact heckin' spots they put their 4 e-core clusters. I have posted the die shots of these CPUs like a day ago, idk if you just chose to ignore them or what lol. Literally everyone use the 4 E-core = 1 P core model... because it's literally true, so idk why you insist on being wrong on this one point.

Pretty crazy tracking Comet -> Tiger -> Alder in the HUB graph. The difference from Comet Lake to Alder Lake is gigantic.

The silicon characteristics of Intel vs TSMC are interesting to see too. Intel has pretty leaky silicon in comparison. Supposedly Intel 4 addresses this with DTCO by removing a dummy gate.

 

[Geddagod]

So ultimately MTL has 9% higher perf at isopower, and that s with an ES chip wich,

I have shown you 4 different power curves from 4 different sources showing there is ~5% improvement in perf/watt with MTL, but honestly, in the grand scheme of things, might not matter much lol.

as i already stated, is eventually overvolted for tests

That was one of the possibilities I stated might be happening (out of the 4 I stated)

in other words the perfs is improved by about 15% at isopower,

No. We have no idea how much it's overvolted by, there's no point of just guessing or making up numbers. Just wait for benches or honestly, just better leaks.

According to AT perf improvement at isopower in this frequency range, 3 to 3.5, is 10% with standard cells and can be pushed to 15% with HP cells.

2 things abt that-
AFAIK 6VT vs 8VT isn't HP vs HD. Both are HP.
They are using arm cores, not RWC in this comparison.

The MLID leak is either not representative of a regular MTL chip (by a large margin) at all, or Intel 4 is a failure. There's really not much in between. The discrepancy between what is expected and what the MLID leak shows is much too large.

 

[mikk]

Intel clarifies MTL is only coming for All-in-One PCs on desktop.

Intel clarifies upcoming 'desktop' Meteor Lake are actually for All-in-One PCs - VideoCardz.com

Intel clarifies confusing statement from the head of its PC division No LGA-1851 for Meteor Lake. In an interview with PCWorld Michelle Johnston Holthaus (Intel Executive Vice President and General Manager of the Client Computing Group) has clearly said that Meteor Lake chips are coming to...

videocardz.com

 

[JoeRambo]

Intel clarifies MTL is only coming for All-in-One PCs on desktop.

Pretty much what was discussed here, coming in NUC style stuff only.
And it makes a lot of sense too, as on proper desktop it would not have chance vs 6Ghz siblings.

 

[Glo.]

Intel clarifies MTL is only coming for All-in-One PCs on desktop.

Intel clarifies upcoming 'desktop' Meteor Lake are actually for All-in-One PCs - VideoCardz.com

Intel clarifies confusing statement from the head of its PC division No LGA-1851 for Meteor Lake. In an interview with PCWorld Michelle Johnston Holthaus (Intel Executive Vice President and General Manager of the Client Computing Group) has clearly said that Meteor Lake chips are coming to...

videocardz.com

Its not coming ONLY for AIOs, but SUCH AS AIO solutions.

Its a key difference.

 

[Glo.]

Pretty much what was discussed here, coming in NUC style stuff only.
And it makes a lot of sense too, as on proper desktop it would not have chance vs 6Ghz siblings.

Nobody is going to stop ASUS, Gigabyte, ASRock, MSI to make desktop form factor boards with soldered MTL-P packages on them.

 

[eek2121]

Nobody is going to stop ASUS, Gigabyte, ASRock, MSI to make desktop form factor boards with soldered MTL-P packages on them.

…except ASUS, Gigabyte, ASRock, and MSI. Why they don’t invest more in smaller form factors is baffling to me. I have a mini PC not much bigger than a Raspberry Pi + case sitting on my desk and it is absolutely fantastic. If I weren’t a gamer, developer, and content creator, it is all I would need. (it can game, just not anything serious)

Not everyone needs or wants to use a laptop or ATX style desktop.

The mini PC has a mount that lets me attach it to a VESA mount. I can attach it to a monitor, and then attach the monitor to a desk mount. The only reason I don’t is I use the mini PC as a headless device.

 

[Glo.]

…except ASUS, Gigabyte, ASRock, and MSI. Why they don’t invest more in smaller form factors is baffling to me.

There is coming very rapidly time, when they will .

Hardware is evolving in the direction where its really tempting for majority of consumers, which will create financial incentives to do that.

 

[igor_kavinski]

MTL isn't even out yet and AMD has the answer to it. Or rather, Lenovo does: https://www.tomshardware.com/news/r...-performance-in-yoga-14s-ryzen-edition-laptop



5% less performance than a 45W Ryzen with 33% less power consumption. Dang, this is going to be a serious thorn in Intel's power efficiency dreams.

 

[eek2121]

MTL isn't even out yet and AMD has the answer to it. Or rather, Lenovo does: https://www.tomshardware.com/news/r...-performance-in-yoga-14s-ryzen-edition-laptop

View attachment 86375

5% less performance than a 45W Ryzen with 33% less power consumption. Dang, this is going to be a serious thorn in Intel's power efficiency dreams.

????

That chart is comparing the 7840hs to a 7940hs and a 6900hs. We have no power consumption numbers for Meteor Lake except from a certain youtuber that self-admittedly knows nothing about technology.

In theory, Intel 4 should be more than competitive with TSMC N4P. We won’t even begin to have an idea or whether it is or not since we are only getting a single Intel 4/3 product. If MTL-P is more/less power efficient than Zen 4, it could come down to the process, architecture or both.

Also note that MTL-P was originally designed as a 28W die. It supposedly scores 7300 in CBr20 at 100W. Even assuming these numbers are accurate, we don’t know how much performance scales with power. Both Intel and AMD push power limits pretty hard for that last 10%. It is entirely possible that MTL-P scores similarly (or beats, or loses) to the 7940hs. We won’t know until we get more details.

 

[Geddagod]

MTL isn't even out yet and AMD has the answer to it

Dang, this is going to be a serious thorn in Intel's power efficiency dreams.

AMD had the answer to it for months now, it's just no one cares. It's not a "thorn for Intel's power efficiency dreams", bcuz Intel has OEMs by the balls lol. AMD's Phoenix "launch" and roll out has been terribly slow and only very recently have we started to see a bunch of phoenix models outside handhelds. And people say Intel paper launches... lol

 

[Geddagod]

In theory, Intel 4 should be more than competitive with TSMC N4P.

I don't think so, why not just competitive

If MTL-P is more/less power efficient than Zen 4, it could come down to the process, architecture or both.

GLC on Intel 7 appears to be as efficient (or marginally less) efficient than Zen 3. If RWC is as efficient as Zen 4, I would be pretty impressed tbh, because Zen 4 increased efficiency from node, but then also increased efficiency through architecture. While I think it's likely RWC has slightly higher architectural perf/watt than GLC as well, I doubt it's anywhere near where Zen 4 got vs Zen 3, and thus if RWC is matching Zen 4 in perf/watt, it's because the node turned out better than expected (or Intel's DTCO team went hard)

We won’t know until we get more details.

Can't wait for Jarrods Tech or HWUB power curve graphs!

 

[Kepler_L2]

I don't think so, why not just competitive

GLC on Intel 7 appears to be as efficient (or marginally less) efficient than Zen 3. If RWC is as efficient as Zen 4, I would be pretty impressed tbh, because Zen 4 increased efficiency from node, but then also increased efficiency through architecture. While I think it's likely RWC has slightly higher architectural perf/watt than GLC as well, I doubt it's anywhere near where Zen 4 got vs Zen 3, and thus if RWC is matching Zen 4 in perf/watt, it's because the node turned out better than expected (or Intel's DTCO team went hard)

Can't wait for Jarrods Tech or HWUB power curve graphs!

HUB stopped doing laptop reviews sadly.