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]

Lol, it s documented by CPU Z themselves, they explained the "reasons" for such a downgrade after they revised the 1.73 version wich was released in december 2016 and "updated" in may 2017 after they noticed the surprisingly good score of the 1800X.

They updated their bench due to the score misrepresenting the actual performance. Zen1 is slower than SKL a bench showing otherwise would not be the intentions of the authors, so they corrected it.

 

[SiliconFly]

-

 

[AcrosTinus]

LMAO, you cannot be serious. Any little credibility you had before that post is now gone man.

The scope of my reply is CPU-Z, other benches in aggregate showed 13%

 

[AcrosTinus]

Possibly. But I just ran the same version with a 13900HX w.r.t 12900K & 13900K and the results were inline with expectations. Might translate well for ARL too. So, I don't think we should label it as totally worthless already.

The most important question we should be asking now is, whether it's altered using photoshop. Is there anyone who can analyze it?


I think so too. But first we need to find out whether it's real or digitally altered.

True

 

[DrMrLordX]

Yes, that is why I say that for comparisons of Intel CPUs it is alright

That is debatable. CPU-z benchmark clearly has something wrong with it. We went through this with Dr. Cutress' 3DPM benchmark when it was found that repeated cache flushing caused the benchmark to heavily favor Intel CPUs with HT on over anything else. It wasn't even a really good tool for comparing Intel CPUs vs other Intel CPUs since the results were so skewed. Arrow Lake-S would probably suffer badly under 3DPM v1.

In any case that screenie might be a fake, and it's an already-sketchy benchmark, so let's not make too much of it.

 

[CouncilorIrissa]

I'm sure the benchmark that fits entirely into the L1 of any modern CPU and practically does not stress the branch predictor is totally representative of all possible workloads.

After all, it's not like pretty much any new CPU microarchitecture presentation contains slides containing words such as "Improved branch prediction". Totally not Apple's most recent presentation. Nor ARM's. Nor AMD's even. Nor Intel's. Companies are just investing transistors into this for lolz.

 

[AcrosTinus]

That is debatable. CPU-z benchmark clearly has something wrong with it. We went through this with Dr. Cutress' 3DPM benchmark when it was found that repeated cache flushing caused the benchmark to heavily favor Intel CPUs with HT on over anything else. It wasn't even a really good tool for comparing Intel CPUs vs other Intel CPUs since the results were so skewed. Arrow Lake-S would probably suffer badly under 3DPM v1.

In any case that screenie might be a fake, and it's an already-sketchy benchmark, so let's not make too much of it.

Fair

 

[AcrosTinus]

I'm sure the benchmark that fits entirely into the L1 of any modern CPU and practically does not stress the branch predictor is totally representative of all possible workloads.

After all, it's not like pretty much any new CPU microarchitecture presentation contains slides containing words such as "Improved branch prediction". Totally not Apple's most recent presentation. Nor ARM's. Nor AMD's even. Nor Intel's. Companies are just investing transistors into this for lolz.

All I'm advocating for is that the term "real world workload" is nothing more than an approximation itself that no benchmark can justify. An aggregate of different benches come closer to representing reality, hence CPU-Z has it's place.

I know CPU-Z is not heavy on the branch predictor but many apps are.

 

[DavidC1]

The scope of my reply is CPU-Z, other benches in aggregate showed 13%

*Sniff* Sniff* I smell WCCFTech...

Wow that's some . . . interesting analysis there!

Hmm, that's a very, very kind way of putting it.

 

[AcrosTinus]

Ignoring everything else I said. How would you like to be on a 4C or 6C CPU today with HT as a ~$100 add on?

I do not indulge in hypotheticals that did not happen. Fact is that 8 core Zen1 was defeated by Coffee Lake as a response (6C/12T). If I remember correctly people were excusing AMD due to better NT performance in that period the same way people excuse Intel for better NT against 7800x3D.

 

[Hail The Brain Slug]

I am right, the bench showed nearly zero IPC gain (CPU-Z), the performance increase in CPU-Z was due to clocks. I think other benches on average showed 10% or 13% increase from Zen3 to Zen4 or is that wrong ?

Until CPU-z corrects the design of its benchmark to respect the CPPC scheduling of Ryzen, the results will be a wildcard because it always forces to core 0, which could be any random quality core boosting to any unknown frequency below or up to max singlethread boost.

Comparing any two samples of Ryzen will give you unpredictable results due to this.

Comparing Ryzen ST results to intel is also disingenuous because it will schedule properly on Intel to the fastest core.

 

[AcrosTinus]

*Sniff* Sniff* I smell WCCFTech...

Hmm, that's a very, very kind way of putting it.

Nice passive aggressive response.
CPU-Z showed no increase in IPC, In aggregate other benches showed 13% increase in IPC from Zen3 to Zen4. If anything I am saying is false, just respond and correct me.

 

[DavidC1]

That is debatable. CPU-z benchmark clearly has something wrong with it.

Let's call spade a spade. CPU-Z is for information, and the benchmark is for laughs. It's the AFV of benchmarking. They decided to get onto a different bandwagon and couldn't be some boring "information provider" anymore.

Nice passive aggressive response.
CPU-Z showed no increase in IPC, In aggregate other benches showed 13% increase in IPC from Zen3 to Zen4. If anything I am saying is false, just respond and correct me.

"Passive aggressive" What is this, Psychology 101 class?

We are ALL telling you CPU-Z sucks. It's a laughingstock. The burden is on YOU to prove that it's a good benchmark.

 

[AcrosTinus]

Until CPU-z corrects the design of its benchmark to respect the CPPC scheduling of Ryzen, the results will be a wildcard because it always forces to core 0, which could be any random quality core boosting to any unknown frequency below or up to max singlethread boost.

Comparing any two samples of Ryzen will give you unpredictable results due to this.

Comparing Ryzen ST results to intel is also disingenuous because it will schedule properly on Intel to the fastest core.

Interesting, where can I read about this scheduling issue in CPU-Z ?

 

[Hail The Brain Slug]

Interesting, where can I read about this scheduling issue in CPU-Z ?

I am telling you as someone who is in possession of multiple generations of Ryzen and Intel systems I have verified this for myself. In fact, I re-verified it just a week or two ago on the latest OS/driver/CPU-z versions to make sure it wasn't fixed.

Attempts to correct the affinity for Ryzen are not fruitful, the benchmark does a sequence of MT then ST, and even if you set affinity during the MT portion it will force it back to core 0 when ST begins.
I have attempted to use Process Lasso to automate the affinity but it doesn't give a good result due to taking some time after the ST bench has begun (and run for some time on a slower core) before it corrects affinity, providing still artificially low scores.

 

[AcrosTinus]

Let's call spade a spade. CPU-Z is for information, and the benchmark is for laughs. It's the AFV of benchmarking. They decided to get onto a different bandwagon and couldn't be some boring "information provider" anymore.

"Passive aggressive" What is this, Psychology 101 class?

We are ALL telling you CPU-Z sucks. It's a laughingstock. The burden is on YOU to prove that it's a good benchmark.

It is passive aggressive. If you want to dismiss a product as a laughingstock, present the evidence, you have to prove the positive otherwise we can just have differing opinions, I don't subscribe to a default negative opinion.

 

[Thunder 57]

It is passive aggressive. If you want to dismiss a product as a laughingstock, present the evidence, you have to prove the positive otherwise we can just have differing opinions, I don't subscribe to a default negative opinion.

Chips & Cheese provided the evidence. You just seem to ignore it or say its "50% crap". Not smart for a new member if they want to gain anyones respect.

 

[AcrosTinus]

So then maybe SPEC?



OK well now you are just being disingenuous because Zen 2 was around just after. Sure CFL (AKA SKL) beat Zen in some areas. But like I said it made 8 core CPU's cost way less than $1000+. What are you trying to prove?

Yes, SPEC is great.

I am trying to prove nothing, I just made the claim that SKL is faster than Zen1 and you want me to somehow praise AMD for competing and making CPUs cheaper after a decade of shame ? , well good for them.

Another hypothetical if their bulldozer succeeded wouldn't we have cheaper CPUs earlier ? The opposite is true as well, due to AMD failure to compete 4C CPUs were sold for a long time.

 

[AcrosTinus]

Chips & Cheese provided the evidence. You just seem to ignore it or say its "50% crap". Not smart for a new member.

I did not say that it is crap, I only call 50% BS on the reasoning that CPU-Z should be totally dismissed due to the instructions mix and how branch light it is. That is all. I think all benches have their place, just put the appropriate weights on it while calculating performance increase.

 

[Hulk]

Lion Cove in ARL will show +14% to a max of +19% average IPC increase over Raptor Cove, with the likely number being about +16%.

Forget about the leaks. Intel already detailed 95% of the changes for Lion Cove when they previewed Lunar Lake. We know about all of the big, important architectural changes. As we have been discussing there will be some minor memory subsystem changes that may results in a few percent increase.

We also know all about the IPC increase of Skymont. The things I'm wondering about now are if ARL will have HT (doubtful in my opinion) and what the all-core frequencies for Lion Cove and Skymont in ARL will turn out to be.

This is actually a nuanced question. For example, the nT frequencies for the 14900K are 5.7GHz and 4.4GHz. The 4.4GHz is kind of realistic but 4.3GHz is a safer day-to-day use bet. 5.7GHz with HT on is ridiculous. You'll need a custom loop, crazy volts, and it'll be over 300W. That's nuts. The "real" specs for the 14900K are more like 5.5/4.3.

So with this in mind if Arrow Lake is specified at 5.4/4.5 all-core and that works with air cooling and reasonable volts and power, say 210W that would be okay. On the other hand if that is 290W and a custom loop then I call BS on the rating. We shall see.

 

[H433x0n]

Holy cow. This thread turned into a mess really fast and devolved into the usual conspiracy drivel.

The performance of ARL-S is more or less known. There are Intel NDA overclocking guides that have already been given to partners and tons of ES samples in the wild. There’s no data to suggest a +20% 1T performance increase. Hype it at your own peril.

 

[poke01]

We should ban CPU-Z results, what a waste of time.

Like what does that benchmark even tell you?

 

[FlameTail]

Forgive me if this question sounds stupid...

So in a desktop motherboard, there is the "Processor" and the "Chipset", as shown in the above diagram.

Is this true for laptop motherboards too? My understanding is that in laptops, the Chipset and Processor are integrated into one SoC. Is that so?

 

[Bouowmx]

Forgive me if this question sounds stupid...

So in a desktop motherboard, there is the "Processor" and the "Chipset", as shown in the above diagram.

Is this true for laptop motherboards too? My understanding is that in laptops, the Chipset and Processor are integrated into one SoC. Is that so?

Except for the HX series, the chipset is moved on-package (regular MCM)

 

[dttprofessor]

Except for the HX series, the chipset is moved on-package (regular MCM)

View attachment 102225

MTL no