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.

 

[inf64]

You're right, the test suite they used is much better than what AMD presented for Zen 5 IPC calculation. If you used that same test suite to calculate IPC for Zen 5, who knows if you actually get a 16% average.

David Huang is running Spec test on Granite Ridge as we speak, so it won't be long now.

 

[ondma]

I don't know why anyone would expect anything else other than marginal ST uplift. Intel 7 is many things, but it allowed Intel to clock their CPUs incredibly high. Finding a large ST uplift was always going to be an uphill battle.

It ia a "new" core, with HT removed, supposedly to give a higher ST boost. That is why I consider the ST gain disappointing. The problem is they are working against a clock speed "offset", so if they had eked out another 5% IPC gain it would have almost doubled the final performance gain.
14% - 8% = 6%
19% - 8% = 11%
The comparison that counts, though, is how the performance stacks up against Zen 5. All the instabilities in 13 and 14 gens are proving they are clocked at unrealistic levels to get higher performance. The real issue though, is that I think ARL will lose to Zen 5 in ST and maybe match it at best in nT. Even worse, is that the next gen for Intel is simply a refresh of ARL, so it wont do much to catch up. Maybe they can at least increase the clock speeds a bit. I see Intel pretty much permanently behind AMD on the desktop unless they hit a home run somewhere. ARL certainly was not it.

As for nT, if the gain is around 15%, i think that is acceptable. I still dont understand the Intel supporters who somehow thought ARL was going to dominate Zen 5 in nT tasks after giving up HT.

 

[H433x0n]

It ia a "new" core, with HT removed, supposedly to give a higher ST boost. That is why I consider the ST gain disappointing. The problem is they are working against a clock speed "offset", so if they had eked out another 5% IPC gain it would have almost doubled the final performance gain.
14% - 8% = 6%
19% - 8% = 11%

I don't get the point of using 14900KS as the bar unless you want to put ARL's performance in the worst light possible. It's not a -KS SKU, it's a -K SKU.

The real issue though, is that I think ARL will lose to Zen 5 in ST and maybe match it at best in nT. Even worse, is that the next gen for Intel is simply a refresh of ARL, so it wont do much to catch up. Maybe they can at least increase the clock speeds a bit. I see Intel pretty much permanently behind AMD on the desktop unless they hit a home run somewhere. ARL certainly was not it.

This is my 14900K in GB6, it's not overclocked or running some insane memory configuration just basic DDR5-6400. This is *really* close to what we've seen from Zen 5 so far. ARL doesn't have to do much to match Zen 5 in 1T performance, the bar it has to clear is to not be a regression.

 

[ondma]

I don't get the point of using 14900KS as the bar unless you want to put ARL's performance in the worst light possible. It's not a -KS SKU, it's a -K SKU.

This is my 14900K in GB6, it's not overclocked or running some insane memory configuration just basic DDR5-6400. This is *really* close to what we've seen from Zen 5 so far. ARL doesn't have to do much to match Zen 5 in 1T performance, the bar it has to clear is to not be a regression.

That is only one benchmark though. I saw CB R23 benchmarks that showed 14900K very close to Zen 4 and Zen 5 with a significant lead over both.

I agree though, we have to wait for final production samples from both camps and look at a wide variety of benchmarks. The thing I dont like though, is that Intel gave up HT, supposedly to enhance ST performance, but the gains being seen are similar to other generations which did not sacrifice HT.

 

[KompuKare]

Even if this QS is representative of final performance, there are still 2 very big unknowns. First is what memory config was used, this can make a significant difference in some benchmarks. The second is power/efficiency. I know it says that it was set to 250W, but we don't know how much it is actually using in CB (my guess would be 250W but we don't know for sure) and how much it uses in idle/ST scenarios.

It ia a "new" core, with HT removed, supposedly to give a higher ST boost. That is why I consider the ST gain disappointing. The problem is they are working against a clock speed "offset", so if they had eked out another 5% IPC gain it would have almost doubled the final performance gain.
14% - 8% = 6%
19% - 8% = 11%
The comparison that counts, though, is how the performance stacks up against Zen 5. All the instabilities in 13 and 14 gens are proving they are clocked at unrealistic levels to get higher performance. The real issue though, is that I think ARL will lose to Zen 5 in ST and maybe match it at best in nT. Even worse, is that the next gen for Intel is simply a refresh of ARL, so it wont do much to catch up. Maybe they can at least increase the clock speeds a bit. I see Intel pretty much permanently behind AMD on the desktop unless they hit a home run somewhere. ARL certainly was not it.

As for nT, if the gain is around 15%, i think that is acceptable. I still dont understand the Intel supporters who somehow thought ARL was going to dominate Zen 5 in nT tasks after giving up HT.

Or maybe not. Win or lose, 1-2% for +40% power is a craziness which has to stop and although Intel certainly started that race (recently), it may take Intel's current woes to put a stop the race too!

 

[Goop_reformed]

Why wait? We don't know final clocks (although the rumors are likely true), final microcode impacts, final price, final availability, final performance of iGPU and NPU, final performance at other power levels, and most relevant to the ST discussion: the final power used in ST usage.

But we do know final clocks right? And the microcode won't have a sizeable impact, as it always has been. And the price will not be competitive at all, have you looked at zen 5 leaked msrp yet?. igpu and npus are still memes. I agree with the last part though, but it's not gonna be any better than zen 5.

 

[dullard]

But we do know final clocks right? And the microcode won't have a sizeable impact, as it always has been. And the price will not be competitive at all, have you looked at zen 5 leaked msrp yet?. igpu and npus are still memes. I agree with the last part though, but it's not gonna be any better than zen 5.

Great, then please list your known Intel clocks for all upcoming Arrow Lake chips here, known final performances here, known Arrow Lake price lists here, and the CPU you will buy in the appropriate thread since you seem to think this is a comparison thread. Then please leave your ignorance about the 2/3rds of computers that use iGPUs and the potential for NPUs at the door.

 

[Goop_reformed]

Great, then please list your known Intel clocks for all upcoming Arrow Lake chips here, known final performances here, known Arrow Lake price lists here, and the CPU you will buy in the appropriate thread since you seem to think this is a comparison thread. Then please leave your ignorance about the 2/3rds of computers that use iGPUs and the potential for NPUs at the door.

No need to get defensive. Performance is already known. Single digit 1t uplift, 15%-20% nt. igpu and npus are still meme because the alternative is very competitive.

 

[dullard]

No need to get defensive. Performance is already known. Single digit 1t uplift, 15%-20% nt. igpu and npus are still meme because the alternative is very competitive.

So we should take your non-answer to mean that you don't know performance, prices, clock speeds, power levels, or how iGPUs and NPUs are necessary?

I'll start you out on something simple. What is the performance, price, and clock speed of the Core UItra 7 255?

 

[Goop_reformed]

So we should take your non-answer to mean that you don't know performance, prices, clock speeds, power levels, or how iGPUs and NPUs are necessary?

I'll start you out on something simple. What is the performance, price, and clock speed of the Core UItra 7 255?

Non answer? I answer to your inquiries directly. For the last time, arrowlake final performance is exactly what intel internal projection has shown.

 

[dullard]

Non answer? I answer to your inquiries directly. For the last time, arrowlake final performance is exactly what intel internal projection has shown.

Directly, what did Intel's internal projection show for the Core UItra 7 255? I missed that in your answer. I do see that you claim "Arrowlake performance is already known at this time", but I can't seem to find that data on the Core UItra 7 255. Heck, I still can't quite figure out the clock speed of that chip, so even if I wanted to apply a generic formula, what clock speed do I use to calculate it?

 

[ondma]

Great, then please list your known Intel clocks for all upcoming Arrow Lake chips here, known final performances here, known Arrow Lake price lists here, and the CPU you will buy in the appropriate thread since you seem to think this is a comparison thread. Then please leave your ignorance about the 2/3rds of computers that use iGPUs and the potential for NPUs at the door.

We have seen leak after leak showing mediocre gains, and the intel group has said, "just wait, its not the final number, there has to be better benchmarks down the road", and every new leak that comes out shows similar gains. Is it "possible" there will be some great revelation that turns things around? Possibly, but I am resigned to the old saying now" If it walks like a duck, looks like a duck, and quacks like a duck, well, it probably IS a duck."

 

[dullard]

We have seen leak after leak showing mediocre gains, and the intel group has said, "just wait, its not the final number, there has to be better benchmarks down the road", and every new leak that comes out shows similar gains. Is it "possible" there will be some great revelation that turns things around? Possibly, but I am resigned to the old saying now" If it walks like a duck, looks like a duck, and quacks like a duck, well, it probably IS a duck."

How much money is that duck? How much energy does that duck use in single threaded waddling? What are the frequency of the duck's feet? How quickly does the duck display images and report answers to questions? I'm mostly interested in the middle ducks in the whole brace of ducks.

 

[Goop_reformed]

How much money is that duck? How much energy does that duck use in single threaded waddling? What are the frequency of the duck's feet? How quickly does the duck display images and report answers to questions? I'm mostly interested in the middle ducks in the whole brace of ducks.

Did you miss the qs tested at 250W?

 

[dullard]

Did you miss the qs tested at 250W?

Not at all. Did you miss the discussion over many forum pages of single thread performance and my specific question about power used by a single thread? For help, see the bolded portion here: "How much energy does that duck use in single threaded waddling?" Which anyone who knows about CPUs is well under the 250 W used for multiple threads.

Hmm, still no direct answer from you.

 

[Goop_reformed]

Not at all. Did you miss the discussion over many forum pages of single thread performance and my specific question about power used by a single thread? Which anyone who knows about CPUs is well under the 250 W used for multiple threads.

Hmm, still no direct answer from you.

No need to get upset. It's just a cpu. You obviously ignore all the leaks, which are ALL legit btw, so far. It's pointless to argue with an individual like this.

 

[dullard]

No need to get upset. It's just a cpu. You obviously ignore all the leaks, which are ALL legit btw, so far. It's pointless to argue with an individual like this.

I'd like to be as informed as you. What are the leaks that I am ignoring regarding the performance, price, and clock speed of the Core UItra 7 255? You did claim that all Arrowlake (sic) performance is known.

 

[SiliconFly]

-

 

[Goop_reformed]

Not even close.

Examples?

 

[DavidC1]

Again, why not just wait until it's review day for both? It's pointless.

SpeedMark, Geekbench, CrossMark, and Cinebench has their own problems and leaks are basically a user submitted benchmark that can have errors easily swinging few low single digit %, which WILL swing opinions again.

Wave #1: People thought Zen 5 is going to be DEFINITELY 32% thus Arrowlake and thus Intel was completely doomed and Pat will be out the door in 5 months.
Wave #2: Computex reveals about uarch of both processors have shown Zen 5 is 32%/2, and Skymont seemed awesome, so now Intel would DOOM AMD.
Wave #3: Certain forumgoers derive that Lion Cove in Arrowlake is somehow radically different and so awesome that it would go from being 14% Lunarlake to basically anywhere from 5 to 10% faster, MLID rabids included.
Wave #4: Now.

WCCFTech should cover the DRAMA in AT forums, so they redeem themselves as at least being a comic relief among the tech press.

 

[DavidC1]

One approach is vastly superior to the other in terms of cost/development cycles/performance/scaling.

I don't really care about the way they reach their performance goals, as long as it meets my expectations.

The 22nm-awesomesauce-trigate process resulted in the 3770K, which was a big disappointment considering how good the predecessor was.

Also you should reserve judgement of how Intel's approach is until products based on them actually arrive. I think a certain 18A based many core processor will do mighty fine.

 

[mikk]

ARL definitely doesn't sound great from all the people who know a bit more including Intel. Wonder why Intel never praised Arrow Lake? Why talking up Panther Lake when it comes much later than ARL. Gelsinger basically said in an interview Lunar Lake and Panther Lake are great and Arrow Lake isn't. You can hear the excitement for LNL and PTL, about ARL nothing. From the beginning people told Lion Cove is a bit underwhelming which seems to turn out true. Also Lion Cove is build for scalability in mind because Royal Core had been cancelled.

 

[majord]

It did seem like he was doing a lot of forward looking to Panther lake during Computex

I'm not even sure Lunar lake excitement is entirely sensible , What is the major drawcard for it? What is intels solution for above 4+4c in 2024?

 

[coercitiv]

What is intels solution for above 4+4c in 2024?

IIRC it's a shrink of MTL on 20A, though not sure it's coming in 2024. Not much is known about it though, people lost interest in their upper mobile lineup the moment we realized it's not Arrow Lake proper.

 

[mikk]

IIRC it's a shrink of MTL on 20A, though not sure it's coming in 2024. Not much is known about it though, people lost interest in their upper mobile lineup the moment we realized it's not Arrow Lake proper.

What are you talking about?

ARL-U 2+8 Intel 3 - rumored MTL refresh
ARL-H 6+8 N3B real Arrow Lake