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.)

 

[Tigerick]

Is it just me or does it really look like LNL is using 2 chiplets in that pic? Thought it was going to be a huge monolithic die tbh

Yep, LNL is a 2-chipset platform with CPU + PCH, similar to Raptor Lake - U, you can refer to this page to have better understanding of LNL with 443 config. That discussion happened half a year ago with @BionicSquash. I didn't know much of improvement is on the GPU, ie GT3. GT3 going to have 50%more Xe cores compared to Meteor Lake, so it is going to be graphics powerhouse. Let's hope Intel ship it before Apple's M4 comes out...

Hmm, these two chips seem closely tie together, anybody know what type of connection Intel is using?? Clearly not DMI

 

[rtxtwt]

wtf....

First Intel Meteor Lake Laptops To Be Very Expensive, Expected To Cost Around $1500 US

The Taiwanese PC industry is expecting Intel's upcoming Meteor Lake CPU-powered laptops to feature a very high price due to their AI tag.

wccftech.com

That's when the entire spectrum of Intel partners will be rolling out their designs but even then, the first laptops aren't going to be cheap. In fact, it is estimated that Intel's Meteor Lake laptops would end up costing above $50,000 NTD which rounds up to around $1500 US. That's a very high price & we have to remember that the Taiwanese pricing does include taxes. But even then, it feels like the 1st Gen Core Ultra CPUs are going to be in a very premium position.

These below were what I knew before, but I never expect situation could be this bad.

Page 59 - Discussion - Leading Edge Foundry Node advances (TSMC, Samsung Foundry, Intel)

Page 59 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

If you are wondering this, some gossip suggested Intel4's manufacturing cost is ~1.8x times of TSMC N6's cost. Source is some comments under leaker 'Golden Pig Upgrade''s Weibo personal page.

Page 862 - Discussion - Intel current and future Lakes & Rapids thread

Page 862 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

What's worth mentioning is Meteorlake in 2H2023 would be minor shipment only, mass shipment would happen in 1H2024.

also I just heard those weibo guys slams MTL power consumption isn't lower even a bit and buggy. I don't know what would happen now and I would still keep conservative about MTL perf.

 

[gdansk]

On the bright side high prices suggest confidence in its performance, right?

 

[dullard]

wtf....

Intel is starting with the high-end Meteor Lake chips in Dec 2023. The rest of the lineup probably comes in 2024. Thus, yes, I would expect the first laptops to be expensive.

Intel Unveils Meteor Lake Architecture: Intel 4 Heralds the Disaggregated Future of Mobile CPUs

www.anandtech.com

Pat Gelsinger, Intel's Chief Executive Officer (CEO), during his opening keynote at Intel Innovation 2023, states that Intel is launching the first Meteor Lake Ultra (high-end) SKU with AI capabilities (NPU) on December 14th....no details on when we can expect other SKUs are set to hit retail channels. Meteor Lake is likely to launch with the top Ultra SKU or at least one that is close to the top of the stack.

 

[S'renne]

Intel is starting with the high-end Meteor Lake chips in Dec 2023. The rest of the lineup probably comes in 2024. Thus, yes, I would expect the first laptops to be expensive.

Intel Unveils Meteor Lake Architecture: Intel 4 Heralds the Disaggregated Future of Mobile CPUs

www.anandtech.com

Lol they just didn't bother checking the latest information from yesterday's event

 

[Panino Manino]

On the bright side high prices suggest confidence in its performance, right?

Don't know about performance but it seems it'll really be a battery champ. AMD regressed a bit with Zen 4 so Intel will have a worthwhile advantage.
For most people more time on battery is more useful than a bit extra performance. IMO

 

[naukkis]

I consider 20% IPC gain a fat bump. They’d need a ~25% IPC increase to break even with a 300mhz clock regression, a ~22-23% IPC increase to break even with a 200mhz clock regression.

I don't understand your math. 300Mhz is only about 6% of Zen4 top bin 1T clock - to break even 6% clock regression they need exactly 6% more frequency as cpu speed is IPC*clock.

 

[H433x0n]

I don't understand your math. 300Mhz is only about 6% of Zen4 top bin 1T clock - to break even 6% clock regression they need exactly 6% more frequency as cpu speed is IPC*clock.

In many applications, they're behind in ST performance. Applications such as Photoshop, V-Ray, WebXprt, GeekBench, etc.

Computerbase has it at 7% less ST performance for 13900K and 10% less ST performance for 13900KS.

 

[inf64]

In many applications, they're behind in ST performance. Applications such as Photoshop, V-Ray, WebXprt, GeekBench, etc.

Computerbase has it at 7% less ST performance for 13900K and 10% less ST performance for 13900KS.

That is true because 13900KS boosts to 6Ghz, while 7950X is maxing out at 5.7Ghz. Let's see Arrow Lake boost to at least 5.5Ghz and get >20% IPC boost

 

[Abwx]

Computerbase has it at 7% less ST performance for 13900K and 10% less ST performance for 13900KS.

For ST Computerbase use CB R20 wich is a best case for Intel and a version of Povray that do not use AVX2 for AMD, only for Intel, otherwise Zen 4 would perform 18% better in this test, third test is CB R15 where there s no advantage for Intel.

Do the maths with the bias from Povray and you ll find that Zen 4 is actually faster in the average of those 3 tests, though that s FP tests actually.

 

[H433x0n]

For ST Computerbase use CB R20 wich is a best case for Intel and a version of Povray that do not use AVX2 for AMD, only for Intel, otherwise Zen 4 would perform 18% better in this test, third test is CB R15 where there s no advantage for Intel.

Do the maths with the bias from Povray and you ll find that Zen 4 is actually faster in the average of those 3 tests, though that s FP tests actually.

I’m not even going to bother with this

That is true because 13900KS boosts to 6Ghz, while 7950X is maxing out at 5.7Ghz. Let's see Arrow Lake boost to at least 5.5Ghz and get >20% IPC boost

Yeah, frequency matters and ARL frequency seems like it’s going to regress much more than Zen 4 -> Zen 5. It’s a bummer for sure.

 

[exquisitechar]

I’m not even going to bother with this

I don't know if the thing he said about POV-Ray is true, but Computerbase's ST test suite is garbage.

 

[Abwx]

I’m not even going to bother with this

Lol, it was proved in this very forum by a member that did tests using the standard Povray distribution and then modding it so that AVX2 is also implemented for AMD.

Actually you re just spreading wishfull thoughts without any technical standing, just talks out of the vaccum that suit yours beliefs, and ignoring actual numbers and facts.

 

[igor_kavinski]

wtf....

Pat has bills to pay and investors to satisfy. Support him!

 

[igor_kavinski]

The only good thing I can say about MTL is that it will help raytracing go mainstream. Hopefully, Intel will encourage developers to do hybrid raytracing so it improves scene quality without tanking the fps. Certain use cases like watching youtube videos or movies may be very good for the laptop's battery but mixed usage will be a mixed bag where I don't expect Intel to gain more than 2 hours of extra battery life over RPL.

 

[H433x0n]

Lol, it was proved in this very forum by a member that did tests using the standard Povray distribution and then modding it so that AVX2 is also implemented for AMD.

Actually you re just spreading wishfull thoughts without any technical standing, just talks out of the vaccum that suit yours beliefs, and ignoring actual numbers and facts.

I’m using your source and taking the results as-is. That’s all I’m doing - it’s super simple. I’m not twisting anything at all.

There’s no point debating this when every time I do there’s always a caveat or an exception where you deem it invalid (unless it shows Zen 4 winning, then it’s the real results of course).

 

[Abwx]

I’m using your source and taking the results as-is. That’s all I’m doing - it’s super simple. I’m not twisting anything at all.

There’s no point debating this when every time I do there’s always a caveat or an exception where you deem it invalid (unless it shows Zen 4 winning, then it’s the real results of course).

But then when the numbers will hiy we can expect some tortuous reasonning to explain the unexplainable,.

When i see that you re saying that RPL has 15% better ST perf than Zen 4 i wonder what is your kool aid, even Computerbase flawed bumbers (because of povray) are below this number, so all you do are inventing numbers that suit your dreams...

CPU-Benchmark: Prozessor-Vergleich

Welche CPU für Games und Apps kaufen? ComputerBase leistet Kaufberatung mit Intel Core i-14000 und AMD Ryzen 7000(X3D) auf Basis von Tests.

www.computerbase.de

 

[H433x0n]

But then when the numbers will hiy we can expect some tortuous reasonning to explain the unexplainable,.

When i see that you re saying that RPL has 15% better ST perf than Zen 4 i wonder what is your kool aid, even Computerbase flawed bumbers (because of povray) are below this number, so all you do are inventing numbers that suit your dreams...

CPU-Benchmark: Prozessor-Vergleich

Welche CPU für Games und Apps kaufen? ComputerBase leistet Kaufberatung mit Intel Core i-14000 und AMD Ryzen 7000(X3D) auf Basis von Tests.

www.computerbase.de

I was referencing 14900K/14900KS when I quoted that 15% number (not 13900K). It may not be 15%, who knows. Right now 13900K and 13900KS 13900KS shows 7% & 10% lead respectively.

If I had to guess RPL-R will show 12% and 15% ST lead for 14900K and 14900KS respectively. We’ll see what results ComputerBase gets when they get a hold of the 14900K and 14900KS.

 

[Abwx]

I was referencing 14900K/14900KS when I quoted that 15% number (not 13900K). It may not be 15%, who knows. Right now 13900K and 13900KS 13900KS shows 7% & 10% lead respectively.

If I had to guess RPL-R will show 12% and 15% ST lead for 14900K and 14900KS respectively. We’ll see what results ComputerBase gets when they get a hold of the 14900K and 14900KS.

Dunno what will be the perfs of the future CPUs and if Intel will still be competitive in ST, for the time be aware that Computerbase ST charts have the limitation i talked about, also that s purely FP.

If we look at other tests that are predominently ST and use INT code the numbers do not turn that favourably to the 13900K.

FI in 7 ZIP ST Zen 4 is 13% faster than RPL, and on WEBXPRT3, Mozilla Kraken RPL does no better than Zen 4 despite a slight frequency advantage, as well as in Geekbench 5.0 ST and Geekbench 5.4 ST, the latter two being FP and INT.

AMD Ryzen 7 7800X3D desktop CPU review: Faster than a Core i9-13900K thanks to 3D V-Cache and only 8 cores

Notebookcheck have tested the new AMD Ryzen 7 7800X3D based on Zen 4 architecture. If offers considerable improvements over Zen3 and impresses primarily thanks to its higher IPC. Its 3D V-Cache leads to a big boost in games.

www.notebookcheck.net

 

[Hulk]

With all of the current talk about Intel nodes I did a little research on what Intel has done with Intel 7 with Alder and Raptor. It will be interesting to see exactly how the Refresh does during objective testing.

Of course at max nT the Intel TDP should be taken with a grain of salt because if those cores are really loaded we know 300+ Watts is possible even at stock settings. And yes, I realize that some people around here have underclocked to get better power results. And yes, I realize that if you back down the power a bit you don't lose much performance and power drops quite a bit. We've been though all this before. My point here is to remind us of where Intel started with Intel 7 with the Golden Cove hybrids when we have the reviews hopefully next month.

Stock	Max nT	Max nT	Max 4T
Settings	P cores	E cores	P cores	TDP		Base P	Base E	TDP
12900K	5.1	3.9	5.2	241W		3.2	2.4	125W
12900KS	5.2	4	5.5	241W		3.4	2.5	150W
13900K*	5.4	4.3	5.8	253W		3.0	2.2	125W
13900KS**	5.4	4.3	6	253W		3.2	2.4	150W
* TechPowerUp noted with sufficient power/cooling the 13900K will run 5.5GHz all-core
** TechPowerUp noted with sufficient power/cooling the 13900KS will run 5.6GHz all-core

 

[Mopetar]

Seems like we're setting a low bar here. And they didn't even demo the product!

It's a deserved bar in my opinion. But I'm not going to turn my nose up at a step in the right direction.

On the bright side high prices suggest confidence in its performance, right?

That's certainly possible. What do Intel laptops with their top-end CPUs normally cost? $1,500 isn't really that expensive for a premium notebook, so unless someone has done some kind of extensive analysis to show these are $$$ more expensive than usual and this increase can't be attributed to other factors, it seems like speculation at best or just clickbait at most likely.

 

[H433x0n]

Dunno what will be the perfs of the future CPUs and if Intel will still be competitive in ST, for the time be aware that Computerbase ST charts have the limitation i talked about, also that s purely FP.

If we look at other tests that are predominently ST and use INT code the numbers do not turn that favourably to the 13900K.

FI in 7 ZIP ST Zen 4 is 13% faster than RPL, and on WEBXPRT3, Mozilla Kraken RPL does no better than Zen 4 despite a slight frequency advantage, as well as in Geekbench 5.0 ST and Geekbench 5.4 ST, the latter two being FP and INT.

AMD Ryzen 7 7800X3D desktop CPU review: Faster than a Core i9-13900K thanks to 3D V-Cache and only 8 cores

Notebookcheck have tested the new AMD Ryzen 7 7800X3D based on Zen 4 architecture. If offers considerable improvements over Zen3 and impresses primarily thanks to its higher IPC. Its 3D V-Cache leads to a big boost in games.

www.notebookcheck.net

Here’s the first 2 results (TPU, Eteknix) on my google search for WebXprt and Kraken benchmarks.

 

[Abwx]

Here’s the first 2 results (TPU, Eteknix) on my google search for WebXprt and Kraken benchmarks.

NBC tests are done at stock setting inclusing the RAM contrary to what you posted, but i guess that you need flawed numbers.

But even then in your link the 7700X perform better than the 7950X in Kraken as well as in Octane, certainly a prove of the tester competence..

I used WEBXPRT3 but your exemple is with WEBXPRT4 wich is now edited by Principled Technologies, you know, that firm that made an advertised and paid comparison between an Intel system with dual channel RAM while the competing AMD solution was granted a single channel...

 

[H433x0n]

NBC tests are done at stock setting inclusing the RAM contrary to what you posted, but i guess that you need flawed numbers.

But even then in your link the 7700X perform better than the 7950X in Kraken as well as in Octane, certainly a prove of the tester competence..

I used WEBXPRT3 but your exemple is with WEBXPRT4 wich is now edited by Principled Technologies, you know, that firm that made an advertised and paid comparison between an Intel system with dual channel RAM while the competing AMD solution was granted a single channel...

This is why I posted “I’m not going to bother with this” initially. I shouldn’t have let myself get sucked into this debate. I won’t make that mistake again.

I’ve provided 3 reputable sources (ComputerBase, TPU & Eteknix) to bolster my point and every example I posted gets shot down as flawed. The only results you’ll accept are the ones that reinforce your existing beliefs. It’s a waste of time.

Edit: Why exactly is it more valid to test at JEDEC memory speeds again? Who’s buying a 7950X and 13900K and running JEDEC DDR5-4800?

If you set both platforms to DDR5-6000 that already gives Zen 4 an advantage since RPL can run DDR5-6400 (realistically it can do DDR5-7200 stable in most cases) without issues.

 

[Timorous]

Edit: Why exactly is it more valid to test at JEDEC memory speeds again? Who’s buying a 7950X and 13900K and running JEDEC DDR5-4800?

People who buy through an OEM rather than do a DIY build, so quite a lot of people. Especially if the machine is productivity focused rather than gaming focused. If you go to Puget for example you get DDR5-5600 in the AMD and Intel configs and there is a note saying actual speed will depend on support so I presume they run it at the CPU manufacturers maximum settings rather than the rated speed of the memory.