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.

 

[Magio]

Apple has made some inroads into Intels market share lead for laptops over the recent years.

I'd argue that while Apple Silicon plays a part there, Windows 11 being complete garbage filled with ads, trackers and bugs with much worse UX/UI than MacOS plays a bigger one. The only semi-conceivable reason to *want* to use Windows today is gaming.

And the fact that OEMs very rarely build laptops of comparable quality to Macbooks (again, without even getting into the actual performance) also doesn't help. For me, Macs are not an option as I run Linux which is a non-starter on Apple Silicon, but believe me when I say it's not easy finding something truly great in the Windows space (especially thin and lights).

 

[cannedlake240]

Clearwater Forest will mop the floor with GNR

Imagine how good it'd be on N3P.. LNL Skymont lp has shown just how inefficient Intel nodes are. Intel 4 is apparently on par with TSMC N6..

 

[AMDK11]

Lion Cove brings absolutely no new ideas, regresses from the predecessor of 3 years ago in some aspects, and is still way bigger than competition, despite having significant process advantage.

Putting a lipstick on a female pig doesn't make it an attractive woman.

It introduces L0-D 48KB and L1-D 192KB alongside L2 3MB. This is something completely new, not only in x86. As for a separate schedule for ALU and a separate schedule for FP, this is nothing new, but a complete overhaul in Intel's P-Core. We will judge whether it is better or worse in some respects after ArrowLake.

LunarLake is not enough to evaluate LionCove in ArrowLake, especially since Intel has officially admitted that both variants of LionCove will differ in several respects.

Edit:
And what new ideas do Mx and Zen5 introduce that are radically new ideas that Intel hasn't had or invented before?

 

[cannedlake240]

Probably why Intel is is no rush to get another P core platform out before Diamond Rapids in H2’26

Lagging Perf/watt on P core Xeons means they're still missing out on deals to Vcache epycs.

 

[cebri1]

Imagine how good it'd be on N3P.. LNL Skymont lp has shown just how inefficient Intel nodes are. Intel 4 is apparently on par with TSMC N6..

Crestmont on Intel 3 is quite efficient. Not so sure about that.

One Xeon 6766E was also about 5% faster than a single Xeon Platinum 8592+ while the Sierra Forest CPU was consuming just 55% the power of the prior Emerald Rapids flagship.

Intel Xeon 6780E / Xeon 6766E 144-Core Performance Benchmarks Review - Phoronix

www.phoronix.com

 

[AcrosTinus]

All this doom and gloom, it's crazy as if people are wanting Intel to fail. Good news arrives doubt it and pick it apart until nothing is left, bad news arrives accept without doubt and proceed to call for Intel's end.

In reality the server parts, according to Phoronix, is competitive and Lunar Lake for thin and lights is good as well. The real Lion Cove will be revealed on Arrow Lake before that time I won't be making any claims about the performance. One thing is sure, AMD has the SIMD advantage on desktop for what it's worth.

 

[cannedlake240]

Crestmont on Intel 3 is quite efficient. Not so sure about that.

Intel Xeon 6780E / Xeon 6766E 144-Core Performance Benchmarks Review - Phoronix

www.phoronix.com

3Ghz all core boost is quite low though, 128C Zen 4C does 3.2Ghz all core at 360W. Maybe Crestmont has the efficiency sweet spot at just below 3Ghz. Apparently Skymont took a latency hit in exchange for higher clocks, so it's plausible... With MTL-U refresh on Intel 3 hitting 5.3Ghz clocks it'll be quite interesting to compare it with Raptor and Meteor-H

 

[cebri1]

3Ghz all core boost is quite low though, 128C Zen 4C does 3.2Ghz all core at 360W. Maybe Crestmont has the efficiency sweet spot at just below 3Ghz. Apparently Skymont took a latency hit in exchange for higher clocks, so it's plausible... With MTL-U refresh on Intel 3 hitting 5.3Ghz clocks it'll be quite interesting to compare it with Raptor and Meteor-H

Bergamo is still so far ahead in IPC. Explains most of the performance difference, after all cresmont is a little bit behind skylake.

 

[poke01]

All this doom and gloom, it's crazy as if people are wanting Intel to fail. Good news arrives doubt it and pick it apart until nothing is left, bad news arrives accept without doubt and proceed to call for Intel's end.

In reality the server parts, according to Phoronix, is competitive and Lunar Lake for thin and lights is good as well. The real Lion Cove will be revealed on Arrow Lake before that time I won't be making any claims about the performance. One thing is sure, AMD has the SIMD advantage on desktop for what it's worth.

I still cannot buy a non-ASUS Strix point laptop. Meanwhile, I can order Lunar Lake from Lenovo, ASUS and MSI. This is a data point based on Australian retailers.

Intel has client in the bag, AMD cannot provide enough laptop Zen 5 dies around the world. As far I’m concerned there are only two players in the laptop space in volume, Intel and Apple.

 

[branch_suggestion]

I still cannot buy a non-ASUS Strix point laptop. Meanwhile, I can order Lunar Lake from Lenovo, ASUS and MSI. This is a data point based on Australian retailers.

Intel has client in the bag, AMD cannot provide enough laptop Zen 5 dies around the world. As far I’m concerned there are only two players in the laptop space in volume, Intel and Apple.

Best to wait for Kraken before declaring it.

 

[511]

If Intel makes the 18A process release on time they will be up against AMD on N3P ... true, but I am GUESSING that Intel will release Q4 2025 at the earliest. Samsung had a he** of a time with GAA. Yields I believe. I just find it hard to believe (and have since Intel came up with the crazy aggressive process roadmap slide) that Intel is going to pull off such a large number of changes into 18A. I guess I just lost faith after the afore mentioned 14++++++ and 10+++++ fiasco .

The DD is there we will see Q325

You may be right about the P core architecture. It seems like the perf/watt of the E cores is impressive in comparison. We will have to re-visit the performance guessing after release and see which one of us was closer .

So Intel 3 is roughly equivalent to N4P and below N3B. This is not different from what I have read other places.

he said density depending on Fin Library as well it is closer to N4P for 2-2 in density and 3-3 library for Intel 3 is closer to N3B
So one Intel 3-3 matches N3 node 3-3 and one Library matches N4P Intel 3 is nicely named
Feels like Samsung is the forgotten one 🤣
but for raw performance it is better
Tsmc Has an Sram Advantage though

So 2024 Turin will be on N4P and therefore be about process equivalent to Granite Rapids Intel 3. Since both designs utilize "P" core only technology, it will be interesting to see how power effective each design will be when placed in the context of being design power limited.

Yes i am feeling RWC is gonna drag it down 🤣

 

[511]

Imagine how good it'd be on N3P.. LNL Skymont lp has shown just how inefficient Intel nodes are. Intel 4 is apparently on par with TSMC N6..

That is totally wrong lol no one would agree to that

 

[511]

Bergamo is still so far ahead in IPC. Explains most of the performance difference, after all cresmont is a little bit behind skylake.

Yes but Clearwater forest will have chadmont to rely upon 288 Chadmont vs 192 5C and we have seen Chadmont are way closer to 5C than Crestmont is to 4C it will be better than 4C outside of AVX-512

 

[DavidC1]

Edit:
And what new ideas do Mx and Zen5 introduce that are radically new ideas that Intel hasn't had or invented before?

Apple M series are worlds ahead in perf/watt. It makes Conroe vs Netburst embarassing in comparison. So despite the later M chips being a small advancement they are already so far ahead it doesn't matter.

Zen 5 has a substantially better branch prediction unit and cluster decode like in Mont which is a departure from Zen 4. Since it's a new idea for them they have bigger headroom to grow.

Lion Cove is more of the same, which has been P core team's motto since Haswell in 2013.

Skymont is kicking ass in efficiency while being able to scale down super low. The branch prediction capability is already exceeding Lion Cove and between Zen 5 and Zen 4. What will Arctic Wolf do? Probably beat the P core in all aspects.

3Ghz all core boost is quite low though, 128C Zen 4C does 3.2Ghz all core at 360W. Maybe Crestmont has the efficiency sweet spot at just below 3Ghz.

Sierra Forest does really well in Cloud and Virtualization workloads and performs almost on par with the best Bergamo. Since that's the original goal for both chips anyway, it is doing it's job, keeping ARM servers out of the market just like Lunarlake is turning WoA into DoA.

 

[DavidC1]

From P6(Pentium PRO) to RedwoodCove, the schedule and execution ports were common to FP and ALU, i.e. FP-ALU. LionCove separates FP from ALU and implements separate/dedicated blocks with execution ports for FP and separate ones for ALU. If no one sees such a fundamental change, I don't know if anyone will ever notice any changes.

This is why the P core team is such a disappointment, even the best IDC that created Conroe/Merom. It took only 30 years to change.

I HIGHLY encourage you to look at the E core and the rapid pace of development and why I think the P core is going to retire just like a cow gets to retire at a slaughterhouse.

I was a big fan of Intel at the time and wanted them to make bigger changes. They barely do. After Sandy Bridge, which still was only 15-20%, but with great efficiency improvements, it was 10% with Haswell, and 10-15% with Skylake.

Meanwhile, after a little struggle from the first OoOE Atom Baytrail, it has been making sweeping changes every major uarch. People make excuses that E core has been getting a lot because it's playing catchup. But every new arch from them, their excuses get thinner and thinner. In fact the excuses don't make sense since the uarch is so divergent.

While Skymont core seems to have been quite enlarged physically, Lunarlake shows that the questionable power efficiency in Grace/Crest has been addressed. It really is an efficient core that also scales enough to be performant.

 

[cebri1]

Intel Core Ultra 9 285KのVrayベンチマーク登場。Core i9-14900Kを26%上回る

Intel Core Ultra 9 285KのVrayベンチマーク登場。Core i9-14900Kを26%上回る Intelは2024年10月24日に次世代デスクトップ向けCPUであるArrow Lake-S世代のC […]

gazlog.jp

9950X is an absolute beast on Vray. Pretty good improvement of the 285K over the 14900KS

 

[AMDK11]

Apple M series are worlds ahead in perf/watt. It makes Conroe vs Netburst embarassing in comparison. So despite the later M chips being a small advancement they are already so far ahead it doesn't matter.

Zen 5 has a substantially better branch prediction unit and cluster decode like in Mont which is a departure from Zen 4. Since it's a new idea for them they have bigger headroom to grow.

Lion Cove is more of the same, which has been P core team's motto since Haswell in 2013.

Skymont is kicking ass in efficiency while being able to scale down super low. The branch prediction capability is already exceeding Lion Cove and between Zen 5 and Zen 4. What will Arctic Wolf do? Probably beat the P core in all aspects.

I'm sorry, but what you wrote in response is gibberish. I asked what new ideas/solutions Mx and Zen5 bring and you told me about energy efficiency and a better predictor. These are completely new solutions and ideas . Massacre.

The predictor idea in Zen5 dates back to the 90's, if you didn't know, and Mx introduces absolutely no new ideas beyond a significant core expansion, including a huge L1-I and L1-D cache. Mx adds more resources like you said about LionCove.

LionCove is moving away from the old Cove approach, but you still don't notice it.

 

[TESKATLIPOKA]

Zenbook S 14 Standard Model 72Wh(Ultra 7 258V, 17W PL1) Link


Zenbook S 16 78Wh(AI 9 HX 370, 28W PL1) Link


Both of them have 32GB RAM, 1TB SSD and the same? OLED display only in different size, so I think a pretty good comparison.

1.) Battery life looks very good, a lot better than Strix.
2.) Lower price than Strix, 1899€ vs 2299€ in my country.
3.) CPU performance is ok, IGP very good
4.) Best perf is at 28W, but It's noisy and they measured temps only with Standard profile.


LNL is pretty good, ok I would have liked 4P+8E cores, but whatever.
My biggest problem with LNL laptops is the price and that they are too thin limiting the cooling as shown above.

P.S. I wonder for how much they want to sell handhelds with LNL?

 

[cannedlake240]

I'm sorry, but what you wrote in response is gibberish. I asked what new ideas/solutions Mx and Zen5 bring and you told me about energy efficiency and a better predictor. These are completely new solutions and ideas . Massacre.

The predictor idea in Zen5 dates back to the 90's, if you didn't know, and Mx introduces absolutely no new ideas beyond a significant core expansion, including a huge L1-I and L1-D cache. Mx adds more resources like you said about LionCove.

LionCove is moving away from the old Cove approach, but you still don't notice it.

Let's see what Panther cove brings

 

[cebri1]

Intel Lunar Lake iGPU analysis - Arc Graphics 140V is faster and more efficient than Radeon 890M

An analysis by Notebookcheck of the Lunar Lake Core Ultra 7 258V coupled with the new Intel Arc Graphics 140V iGPU.

www.notebookcheck.net

 

[exquisitechar]

Well, it looks like Lunar Lake laptops are the ones to get if you don’t want to deal with Apple. Intel did a good job. Good battery life and good overall performance. AMD dropped the ball with the disappointing Strix and the only reasons to consider buying it instead are its niche of high MT performance and maybe AMD’s GPU drivers being pretty solid compared to the competition. Qualcomm’s offerings are mostly just dead in the water now, there’s little reason to deal with WoA issues or the horrid GPU drivers when Lunar Lake exists.

 

[Magio]

Really hope we'll soon have a larger offering of LNL designs because the chips themselves are basically everything I wanted out of my next upgrade.

 

[SiliconFly]

.

 

[cannedlake240]

Intel 4 is closer to N3 (than N5 or N6).

Huang guy's LNL tests say otherwise. He compared Crestmont on the CPU tile vs Crestmont on the N6 soc tile, both showed similar perf at the same power(cores without package). Idk could be that soc is using an ultra efficient variant of N6.

Actually, it's kinda fashionable to bad mouth Intel these days

Well to be fair, they are currently experiencing probably the most severe brain drain in years, talent both new and old, lifers are leaving for other companies. They had hoped for a turnaround when Pat came back but instead the stock got pummeled even more mainly because of foundry expenses, Intel's pay already wasn't great vs. likes of alphabet, apple

 

[NTMBK]

Best to wait for Kraken before declaring it.

I've been hearing this since the Llano days. "Wait for Trinity, that's the real competitive laptop part!"

AMD just don't have the connections and the volume, and I don't know why at this point. Their APUs have been excellent for years now, but they're still losing out to Intel left right and center. My only conclusion is that Intel can out manufacture them with the sheer capacity of their fabs, while AMD has to fight Nvidia and Apple for space.