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

[Tigerick]

Wildcat Lake (WCL) Preliminary Specs

Intel Wildcat Lake (WCL) is upcoming mobile SoC replacing ADL-N. WCL consists of 2 tiles: compute tile and PCD tile. It is true single die consists of CPU, GPU and NPU that is fabbed by 18-A process. Last time I checked, PCD tile is fabbed by TSMC N6 process. They are connected through UCIe, not D2D; a first from Intel. Expecting launching in Q2/Computex 2026. In case people don't remember AlderLake-N, I have created a table below to compare the detail specs of ADL-N and WCL. Just for fun, I am throwing LNL and upcoming Mediatek D9500 SoC.

		Intel Alder Lake - N	Intel Wildcat Lake	Intel Lunar Lake		Mediatek D9500
Launch Date		Q1-2023	Q2-2026 ?	Q3-2024		Q3-2025
Model		Intel N300	?	Core Ultra 7 268V		Dimensity 9500 5G
Dies		2	2	2		1
Node		Intel 7 + ?	Intel 18-A + TSMC N6	TSMC N3B + N6		TSMC N3P
CPU		8 E-cores	2 P-core + 4 LP E-cores	4 P-core + 4 LP E-cores		C1 1+3+4
Threads		8	6	8		8
Max Clock		3.8 GHz	?	5 GHz
L3 Cache		6 MB	?	12 MB
TDP		7 W	Fanless ?	17 W		Fanless
Memory		64-bit LPDDR5-4800	64-bit LPDDR5-6800 ?	128-bit LPDDR5X-8533		64-bit LPDDR5X-10667
Size		16 GB	?	32 GB		24 GB ?
Bandwidth			~ 55 GB/s	136 GB/s		85.6 GB/s
GPU		UHD Graphics		Arc 140V		G1 Ultra
EU / Xe		32 EU	2 Xe	8 Xe		12
Max Clock		1.25 GHz		2 GHz
NPU		NA	18 TOPS	48 TOPS		100 TOPS ?

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.

 

[AcrosTinus]

My new first follow up.

"You say the results were unexpected. What was Intel expecting exactly?"

Then sit back and watch the dance.

It could be that the product was rushed and the TTM was priority here instead of delaying. After that use the user and reviewers as Beta testers, digest the information, craft the optimization and new microcode/firmware. Voilà, you have optimized your CPU while cutting down on some internal processes AT THE COST OF YOUR REPUTATION.

Intel man, you better get a x3D equivalent or brute force IPC gain to be competitive or it might soon be over in the DIY space...

 

[Kocicak]

BTW, good advice some people here regarding my 3rd degraded part. I'm going to ask for a refund. I don't want another defective part as I believe they are all the same (defective).

They are not defective physically, they are sold with defective frequencies. You can fix that, if you want. It is up to you.

Now I run my new mediocre 14900K with 5/4 GHz frequency limits, HT off and it performs really nicely, for example, consumes just 55W while gaming and it is cool as well.

 

[511]

They are not defective physically, they are sold with defective frequencies. You can fix that, if you want. It is up to you.

Now I run my new mediocre 14900K with 5/4 GHz frequency limits, HT off and it performs really nicely, for example, consumes just 55W while gaming and it is cool as well.

View attachment 111391

In the end you are not getting what you paid for ARL seems to have solved the issue of degradation and stuff but it will take time to tell whether the Microcode fixed it or not i know some people who have been running their processors fine some are not

 

[static shock]

ARL and Zen5 should be faster. Lion Cove underperforms versus core ultra 100 p-core in IPC metric.

 

[511]

Lion cove is the biggest waste of sand in ARL silicon

 

[OneEng2]

Perhaps....still can you imagine how pissed people here would be if ARL had only Skymont cores in it?

 

[Hulk]

It could be that the product was rushed and the TTM was priority here instead of delaying. After that use the user and reviewers as Beta testers, digest the information, craft the optimization and new microcode/firmware. Voilà, you have optimized your CPU while cutting down on some internal processes AT THE COST OF YOUR REPUTATION.

Intel man, you better get a x3D equivalent or brute force IPC gain to be competitive or it might soon be over in the DIY space...

Okay, let's role play.

Are the benches produced by reviewers in line with Intel's press numbers? If so, then what was unexpected? If not, what exactly was different and what caused this?

 

[Hulk]

They are not defective physically, they are sold with defective frequencies. You can fix that, if you want. It is up to you.

Now I run my new mediocre 14900K with 5/4 GHz frequency limits, HT off and it performs really nicely, for example, consumes just 55W while gaming and it is cool as well.

View attachment 111391

Okay let's go down that rabbit hole.

The CPU is advertised to run 5.6GHz all-core, 6GHz single core but burns out if you do that. Either the part is defective because it can't hold the rating, or the rating is defective because it over rates the part. Either way the package/part is defective. It doesn't do what the manufacturer claims it can do.

It's word play. I understand your position. There is nothing inherently wrong with the physical construction of the CPU, Intel rates it too high and therefore it fails. In a court of law I think the jury would side with the plaintiff, meaning us.

 

[Hulk]

Perhaps....still can you imagine how pissed people here would be if ARL had only Skymont cores in it?

But then again, if ARL was 4+24 the ST performance would be basically the same, lightly threaded apps would have performed basically the same because of the narrow gap performance-wise between Lion Cove and Skymont, but Intel would have had a massive feather in their cap because MT performance in apps that could have utilized all threads would have been through the roof. Over 2800 CB R24 MT.

Go for a win (MT) and a loss (ST) rather than a tie (MT) and loss (ST). They could have also said they focuses on Skymont and MT performance because that is where software is heading. They will work on ST in the next iteration. It would at least have made some sense.

 

[511]

But then again, if ARL was 4+24 the ST performance would be basically the same, lightly threaded apps would have performed basically the same because of the narrow gap performance-wise between Lion Cove and Skymont, but Intel would have had a massive feather in their cap because MT performance in apps that could have utilized all threads would have been through the roof. Over 2800 CB R24 MT.

Go for a win (MT) and a loss (ST) rather than a tie (MT) and loss (ST). They could have also said they focuses on Skymont and MT performance because that is where software is heading. They will work on ST in the next iteration. It would at least have made some sense.

If they are going to loose ST better to go 32 SKT wipe everything in MT

 

[static shock]

Skymont is a marvel.

 

[511]

Okay let's go down that rabbit hole.

The CPU is advertised to run 5.6GHz all-core, 6GHz single core but burns out if you do that. Either the part is defective because it can't hold the rating, or the rating is defective because it over rates the part. Either way the package/part is defective. It doesn't do what the manufacturer claims it can do.

I think it was 5.4-5.5 ghz not 5.6 ghz
You are correct though you are cheating with the customer if your product doesn't behave as advertised

 

[511]

I don't know a single person on the internet who would say no to a 24C or 32C skymont chip

 

[Hulk]

I think it was 5.4-5.5 ghz not 5.6 ghz
You are correct though you are cheating with the customer if your product doesn't behave as advertised

It's a detail but kind of important because it show how loose Intel played it with frequency.

Ark shows 5.6GHz for max P core boost, plus another 100MHz for TVB, which was so relaxed that it would always engage if power was there and temps were below throttling, which would mean 5.7GHz nT.
5.7 is instanity for Raptor Lake and long term viability

Supported by this review.

Intel Core i9-14900K Review - Reaching for the Performance Crown

Core i9-14900K is Intel's new flagship with clock speeds of up to 6 GHz. It's actually clocked even higher than the 13900KS, thanks to an extra 100 MHz when more than two cores are active. Our review confirms that Raptor Lake Refresh is amazing for both applications and gaming, if you can live...

www.techpowerup.com

 

[Hulk]

I don't know a single person on the internet who would say no to a 24C or 32C skymont chip

Perhaps 28+2 would be the sweet spot?

First thing we'd (or at least I would like) need to see is head-to-head Skymont vs Lion Cove, which is hard to do since we're not able to shut down all Lion Coves in the compute complex.

 

[gdansk]

I don't know a single person on the internet who would say no to a 24C or 32C skymont chip

Nice to meet you.

 

[Hulk]

Nice to meet you.

Very good! Short, sweet, funny, and makes the point. This is why I still spend time here. We can disagree and still have fun.

 

[511]

Nice to meet you.

Nice to meet you too what is your dream ARL config ?

 

[gdansk]

Nice to meet you too what is your dream ARL config ?

I would like good 1T and gaming performance. So at least 6P cores.
Unless it's a very cheap part. Then it's forgivable - like a hypothetical Arrow Lake N.

 

[511]

I would like good 1T and gaming performance. So at least 6P cores.
Unless it's a very cheap part. Then it's forgivable - like a hypothetical Arrow Lake N.

I can give you 24P cores with 4.9-5.0ghz all core
Good gaming okay ST good multi

 

[511]

I can't believe this

I really want to know what MLID Smokes

 

[SiliconFly]

.

 

[MoistOintment]

My new first follow up.

"You say the results were unexpected. What was Intel expecting exactly?"

Then sit back and watch the dance.

If we take what he's saying at face value, then the performance of the CPU is below the expectations of the engineers that designed it, despite matching the results seen by the team that ran internal benchmarks shortly before it was released.

He certainly could be doing PR damage control. But if we give the benefit of the doubt for a minute and assume he's telling the truth, there's certainly no contradiction here because the team that would do a 30 game internal benchmark shortly before the product unveiling would be a totally different team than the ones who designed the CPU.

 

[Saylick]

“Exclusive:-#Panther Lake is 35% more Efficient than #Lunar Lake.”

Looks like it’s totally made up. He posts way too much unverifiable/cooked-up stuff & lots of garbage. Wouldn’t trust him unless the news comes from a reliable source.

Prakhar the Intel fanboy is going to fanboy. More news at 11.

Here’s a tweet in the same thread where he says he doesn’t actually know if Xe3 is actually 40% faster than Xe2, contradicting himself:

https://twitter.com/x/status/1788644719519547543

Beyond_FPS sarcastically says that there’s a worldwide conspiracy to get people to buy Zen 5 and the obvious joke flies over Prakhar’s head…


He also has an alt Xitter account whose entire purpose is to promote Intel, x86, and Windows.

x.com

x.com

 

[mikk]

His anger is justified. Another 3 to 4 weeks wait is kinda painful.


+35% efficiency isn't easy. In fact it close to impossible unless they pull out some major new gimmicks.

From N3B to 18A, the ppw gain (power, not density) isn't that significant I think. It's like 10% or maybe even upto 20% due to BSPDN. 35%? Thats just way off. In fact, I'm getting the feeling that it's gonna be hard for PTL-H to even match LNL efficiency (let alone +35%).

Edit: If PTL-H can comfortably surpass LNL in efficiency, it places Intel in a whole new level. A direct competitor to Apple leaving AMD in the dust. So, this +35% efficiency gain rumor should be taken with a tiny truck load of salt.

It is very very easy for PTL with 8 more E-cores, in fact +35% isn't enough to even match Strix Point at 25W, not even close.