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.

 

[DavidC1]

Let's not be hyperbolic here. Crestmont isn't at least 5%. That imples potentially much as 10%. No, we're going to get 3-4%, which is almost margin of error difference since we have crappy reviewers, and no desktop parts meaning nothing ISO-comparisons which is the cardinal rule for benchmarking.

Also, we got 5% in Raptorlake so that's not zero. Arrowlake got 10% too, despite the shoddy uncore.

The disappointment is that Pantherlake isn't getting anything, because Novalake, the successor after that might also end up <10%. If Novalake is 6.2GHz+ then I guess he will be right in the long run with a N2 2027 core getting 0%*, so what am I doing typing this lol?

*So I guess getting 50% margin on a future product means investing zero in uarch and just clocks?

So single core isn’t everything

It's basically the sole reason why we care about CPUs. And time after time it's demonstrated a wider, lower clocked core is better in just about every aspects, which is entirely single core domain.

It's also because we implicitly known that single thread advancements are the hardest to advance, thus is rewarded most handsomely.

FP, AI, multi cores, they are in the accelerator domain, where just adding more generally helps a ton.

 

[adroc_thurston]

lower clocked core

Everyone is shipping 5G+ next year.

 

[poke01]

Everyone is shipping 5G+ next year.

What happens when a 10-11 wide core with M4 like IPC, reaches 5GHz or even 5.5GHz for the laptop variants?

A blood bath for Intel client laptop.

 

[DavidC1]

Everyone is shipping 5G+ next year.

You constantly prove my point about x86 vendors being retarded, and why they need a complete rethink.

What happens when a 10-11 wide core with M4 like IPC, reaches 5GHz or even 5.5GHz for the laptop variants?

A blood bath for Intel client laptop.

Ah, finally someone that understands.

Not only that, they already outclass desktop parts that should have much, much less thermal density with heatsinks that weigh more than some laptops.

Most importantly, 10-11 pipeline stages.

 

[adroc_thurston]

What happens when a 10-11 wide core with M4 like IPC, reaches 5GHz or even 5.5GHz for the laptop variants

Means we need to get to 6.5.

You constantly prove my point about x86 vendors being retarded, and why they need a complete rethink.

?
You say clocks bad, yet everyone is cranking them higher and higher.
Where's the magic IPC juice? Why is no one shipping Royal Cores at 3.5GHz?

 

[511]

@Kepler_L2 you said NVL is PL1=150W and PL2 is 320W and t=28s does the performance estimate adhere to these setting for the 1.6X gains?

 

[AcrosTinus]

A follow up to the IGPU case, after buying a 285K and trying the encoder with handbrake it shows the same speeds, not comparable to the Raptor Lake IGPU. Slowly I am believing that Intel nerfed the decoder/encoder for increased 3D performance...

Can anyone validate with handbrake using anything from AlderLake++ IGPU vs. Intel Graphics on 200S ?

 

[511]

A follow up to the IGPU case, after buying a 285K and trying the encoder with handbrake it shows the same speeds, not comparable to the Raptor Lake IGPU. Slowly I am believing that Intel nerfed the decoder/encoder for increased 3D performance...

Can anyone validate with handbrake using anything from AlderLake++ IGPU vs. Intel Graphics on 200S ?

i have MTL CPU with the same Encoder/Decoder can you tell me what to test?

 

[AcrosTinus]

i have MTL CPU with the same Encoder/Decoder can you tell me what to test?

Thank you , I would use any generic phone video with 6 to 20K bitrate then try using H.265 or H.264 at file framerate and the quality preset from handbrake, the UHD770 at ICQ 19 would still deliver 180 frames with the Quick Sync encodes. The Intel Graphics on the other hand 20 to 45.

 

[511]

Thank you , I would use any generic phone video with 6 to 20K bitrate then try using H.265 or H.264 at file framerate and the quality preset from handbrake, the UHD770 at ICQ 19 would still deliver 180 frames with the Quick Sync encodes. The Intel Graphics on the other hand 20 to 45.

specific Handbrake version?

 

[AcrosTinus]

specific Handbrake version?

The Version 1.9.2

 

[511]

Thank you , I would use any generic phone video with 6 to 20K bitrate then try using H.265 or H.264 at file framerate and the quality preset from handbrake, the UHD770 at ICQ 19 would still deliver 180 frames with the Quick Sync encodes. The Intel Graphics on the other hand 20 to 45.

Original file size

 

[AcrosTinus]

View attachment 126691

Original file size
View attachment 126692

That means that my whole setup is borked, MTL has exactly the same IGPU as 200S, it isn't a version with more slices, right ?

 

[Io Magnesso]

Means we need to get to 6.5.

?
You say clocks bad, yet everyone is cranking them higher and higher.
Where's the magic IPC juice? Why is no one shipping Royal Cores at 3.5GHz?

Right now, increasing the clock is the quickest way to improve performance.
First of all, it would be better to receive the reality that the ARMs are also increasing the clock frequency.

 

[511]

That means that my whole setup is borked, MTL has exactly the same IGPU as 200S, it isn't a version with more slices, right ?

UHD 770 doesn't support AV1

GitHub - intel/media-driver: Intel Graphics Media Driver to support hardware decode, encode and video processing.

Intel Graphics Media Driver to support hardware decode, encode and video processing. - intel/media-driver

github.com

Media Capabilities Supported by Intel Hardware

Learn about media capabilities of various generations of Intel hardware.

www.intel.com

 

[511]

Right now, increasing the clock is the quickest way to improve performance.
First of all, it would be better to receive the reality that the ARMs are also increasing the clock frequency.

@adroc_thurston is right Apple has achieving majority of ST Improvements from clocks the IPC hasn't improved much compared to Intel/AMD IPC improvements

 

[igor_kavinski]

That means that my whole setup is borked, MTL has exactly the same IGPU as 200S, it isn't a version with more slices, right ?

https://www.reddit.com/r/PleX/comments/1lh5bl0/hevc_encoding_testing_w_core_ultra_igpus_4k_to_4k

It's faster at least in HEVC. Maybe there is some BIOS option issue or your mobo's BIOS is doing something stupid to hold back the iGPU?

 

[Io Magnesso]

View attachment 126636

https://m.weibo.cn/detail/5183934787748234

Wonderful
And this person a fool who previously wanted a Nova Lake made with TSMC N2 on X.

https://twitter.com/x/status/1938993753328586941

No matter how much it's true that Panther Lake hasn't changed, it's ridiculous to say such a remark.😅 It's ridiculous🤪
Just using the TSMC N2 will get better That's the kind of magic that is said in this tweet that doesn't have a sense of reality.

 

[AcrosTinus]

https://www.reddit.com/r/PleX/comments/1lh5bl0/hevc_encoding_testing_w_core_ultra_igpus_4k_to_4k

It's faster at least in HEVC. Maybe there is some BIOS option issue or your mobo's BIOS is doing something stupid to hold back the iGPU?

I have a feeling that MSI's 200S boost BIOS might be the issue here, I will try reverting and encoding a video (H.265) later, currently the IGPU in 200S is slower than the UHD770 in my 13700K and even G7 from 11th gen...

 

[511]

I have a feeling that MSI's 200S boost BIOS might be the issue here, I will try reverting and encoding a video (H.265) later, currently the IGPU in 200S is slower than the UHD770 in my 13700K and even G7 from 11th gen...

lol btw the Xe core slice doesn't matter cause the encoder and decoder are on N6 SoC tile which is shared between ARL/MTL

 

[511]

lol btw the Xe core slice doesn't matter cause the encoder and decoder are on N6 SoC tile which is shared between ARL/MTL

this means that you can run Quick Sync on F SKU and they can have display output as well

 

[Io Magnesso]

Wonderful
And this person a fool who previously wanted a Nova Lake made with TSMC N2 on X.

No matter how much it's true that Panther Lake hasn't changed, it's ridiculous to say such a remark.😅 It's ridiculous🤪
Just using the TSMC N2 will get better That's the kind of magic that is said in this tweet that doesn't have a sense of reality.

After all, this person is ridiculous
He's a painful person who is just self-conceived as a smart person who can see things.
A truly clever person can't expect it even if Intel uses N2.

 

[511]

After all, this person is ridiculous
He's a painful person who is just self-conceived as a smart person who can see things.
A truly clever person can't expect it even if Intel uses N2.

well Nova Lake has another improvements that simply can't be seen with ST/MT improvements also whoever is spreading rumor about N2X next year ain't gonna happen

 

[Io Magnesso]

well Nova Lake has another improvements that simply can't be seen with ST/MT improvements

I don't give a damn
As with NVL/Zen6 at the moment, only the number of cores is known. Too little information

 

[Io Magnesso]

I don't give a damn
As with NVL/Zen6 at the moment, only the number of cores is known. Too little information

I didn't think he was such a painful guy
David Huang is much better than that guy.