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.

 

[poke01]

M4 is going to massacre competition. Apple vs x86 right now is a settled debate.

Good thing Intel's main competition is AMD and Qualcomm. Where Intel has the lead in ST and battery life.

 

[511]

M4 is going to massacre competition. Apple vs x86 right now is a settled debate.

Yes no doubt about ST perf and efficiency in ST and but in their wallet garden there is no way Intel/AMD could achieve Apple level performance until major changes in their arch

 

[jdubs03]

Are we expecting 288V to get close to M3 in MT?
It uses 30 watts as well in Cinebench R23. Intel's Skymont E-cores are more powerful than Apple's so it might.
View attachment 107129
View attachment 107130

I really don’t see why not with the p-core and vastly upgraded e-core performance. But maybe I’m just in the optimistic camp there.

 

[naukkis]

Expecting too much in reference to what?

Mt scaling from st performance when st is boosted to full tdp and beyond.

 

[naukkis]

I really don’t see why not with the p-core and vastly upgraded e-core performance. But maybe I’m just in the optimistic camp there.

Intel Lunar Lake scaling in those intel slides is pretty flat after 17w. After all there is only 4 e-cores - p cores are only better than before - still awfully inefficient.

 

[mikk]

Are we expecting 288V to get close to M3 in MT?
It uses 30 watts as well in Cinebench R23. Intel's Skymont E-cores are more powerful than Apple's so it might.
View attachment 107129
View attachment 107130

No this is not expected I would say. The scaling from 17W to 30W was about 25% in Cinebench R23, I would expect a similar gain in CB R24 which makes it about 600 points at 30W unless the 479 score at 17W is way off but doesn't look that wrong to me. Sure other devices with a better cooling and lower temperatures might perform slightly better.

 

[jdubs03]

So if we go by these slides below, the 288V looks alright on multithreaded workloads. I’m sure that the data could be reverse engineered to get a decent approximation.

The HX 370 was in the Asus Zenbook S16 which gets 1008 points in CB24 and is shown the image below (based on computerbase data). Unfortunately, the 165H is on an Intel development device so we can’t really get a beat on what that performance level was.


Obviously the below image is giving some lower values. In fact the U7 155H seems to score a good bit better with similar wattages reported. So small sample size and probably bad cooling affect the results there.

But based on everything above, idk why a score over 700 isn’t possible. Considering they report only about a 6 percent lower score for the 288V at 23W compared to the U7 165H. Eyeballing it at 28W, looks like a 12% or so deficit to the 165H, and a similar 12ish% deficit to the HX 370 at 33W. Going by that, the score would be 880. It certainly doesn’t look more than a 20% deficit, which would put it below 800. And when considering the M3 in the MBP, which scores 714; clearly they’re saying that LNL can equal it at the same power level. And at 33W it looks like it’s halfway between 714 and 1008, or around 860.

 

[mikk]

The Dell can't hold sustained 17W package power because of temperature throttling by the looks of it.

 

[cebri1]

The Dell can't hold sustained 17W package power because of temperature throttling by the looks of it.

“do not use for benchmarks”

 

[igor_kavinski]

“do not use for benchmarks”

Either the fans are not being engaged or they purposely used just enough thermal solution to let it boot to Windows desktop and maybe run a few apps.

 

[cebri1]

Either the fans are not being engaged or they purposely used just enough thermal solution to let it boot to Windows desktop and maybe run a few apps.

it could be many things, that’s why I would disregard anything coming out of that video. Independent reviews are just 2 weeks away.

 

[whoshere]

Aida64 Cache and Memory Benchmark Intel Core i7 265KF + DDR5 6400c32:

Latency looks a tad high, L2 bandwidth is a tad low, otherwise it's great.

 

[MS_AT]

Aida64 Cache and Memory Benchmark Intel Core i7 265KF + DDR5 6400c32:

Latency looks a tad high, L2 bandwidth is a tad low, otherwise it's great.

And how Aida accounts for L0? Is it bundled with L1 or any other permutation?

 

[511]

Aida64 Cache and Memory Benchmark Intel Core i7 265KF + DDR5 6400c32:

Latency looks a tad high, L2 bandwidth is a tad low, otherwise it's great.

9700x/9900x for reference why do i feel like the soc is screwing memory latency a bit

 

[511]

And how Aida accounts for L0? Is it bundled with L1 or any other permutation?

The CPUID is un changed from software perspective to avoid potential issues

 

[whoshere]

And some juicy Lunar Lake Intel Core i9 269V shots.

Idle power consumption, less than 0.2W, is something AMD should be envious of. My 7840HS "idles" at 3.5W.

Desktop AMD CPUs? 20-45W depending on your memory settings. 10-16W if you're running at JEDEC settings (which almost no one does).

 

[511]

And some juicy Lunar Lake Intel Core i9 269V shots.

Idle power consumption, less than 0.2W, is something AMD should be envious of. My 7840HS "idles" at 3.5W.

Desktop AMD CPUs? 20-40W depending on your memory settings. 10-16W if you're running at JEDEC settings (which almost no one does).

They were not kidding idling at 0.17W - 0.5W is insane

 

[majord]

And some juicy Lunar Lake Intel Core i9 269V shots.

Idle power consumption, less than 0.2W, is something AMD should be envious of. My 7840HS "idles" at 3.5W.

Desktop AMD CPUs? 20-45W depending on your memory settings. 10-16W if you're running at JEDEC settings (which almost no one does).

I think something's up with your power plan or BIOS if your 7840 is idling that high! , even if it is a HS class.

6850U here idles around the 0.5 - 0.6W mark on balanced. (package power) possibly less when he screen's off , i'd have to check (which also may be why the LL there has a very low logged min)

 

[MS_AT]

The CPUID is un changed from software perspective to avoid potential issues

So how are caches reported? L0 is omitted from the report and the CPU reports it has only standard L1-L3 setup?

 

[Det0x]

9700x/9900x for reference why do i feel like the soc is screwing memory latency a bit

9950X when tuned, lower latency is possible, even with the soc limitation

The cache latency on that 265KF seem alittle on the highside tho, maybe p&e core bug on aida (?)

 

[511]

So how are caches reported? L0 is omitted from the report and the CPU reports it has only standard L1-L3 setup?

https://twitter.com/x/status/1828656628645695896

 

[511]

I think something's up with your power plan or BIOS if your 7840 is idling that high! , even if it is a HS class.

I agree should be around 2-3W range but knowing windows

6850U here idles around the 0.5 - 0.6W mark on balanced. (package power) possibly less when he screen's off , i'd have to check (which also may be why the LL there has a very low logged min)

Is it the minimum or it can even below LNL has tiles as well it shows superiority of the packing tech they possess

 

[msj10]

Aida64 Cache and Memory Benchmark Intel Core i7 265KF + DDR5 6400c32:

Latency looks a tad high, L2 bandwidth is a tad low, otherwise it's great.

mem latency is 10 ns worse than RPL. I thought it would be much worse than that based on MTL.

 

[Josh128]

Yeah those are very disappointing. On both ST and MT. My assumption was it was not plugged in. But still that MT score is awful, seems like it should easily be higher, I was expecting around 700-800.

ST is not disappointing at all here. @5GHz it is beating Strix @5.1 by 8-10%. Its beating the newest, fastest x86 core available by 10%, what more do you want? MT is nothing special, but if this is at 17W, its ~passable. Strix gets 954 @28W with 12 cores.

 

[9949asd]

Aida64 Cache and Memory Benchmark Intel Core i7 265KF + DDR5 6400c32:

Latency looks a tad high, L2 bandwidth is a tad low, otherwise it's great.

Still something not right. L2 are too slow. And for ARL everyone can use CUDIMM xmp 9200.