Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads

[Tigerick]

Wildcat Lake (WCL) Specs

Intel Wildcat Lake (WCL) is upcoming mobile SoC replacing Raptor Lake-U. 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 Q1 2026.

		Intel Raptor Lake U	Intel Wildcat Lake 15W?	Intel Lunar Lake		Intel Panther Lake 4+0+4
Launch Date		Q1-2024	Q2-2026	Q3-2024		Q1-2026
Model		Intel 150U	Intel Core 7	Core Ultra 7 268V		Core Ultra 7 365
Dies		2	2	2		3
Node		Intel 7 + ?	Intel 18-A + TSMC N6	TSMC N3B + N6		Intel 18-A + Intel 3 + TSMC N6
CPU		2 P-core + 8 E-cores	2 P-core + 4 LP E-cores	4 P-core + 4 LP E-cores		4 P-core + 4 LP E-cores
Threads		12	6	8		8
Max Clock		5.4 GHz	?	5 GHz		4.8 GHz
L3 Cache		12 MB		12 MB		12 MB
TDP		15 - 55 W	15 W ?	17 - 37 W		25 - 55 W
Memory		128-bit LPDDR5-5200	64-bit LPDDR5	128-bit LPDDR5x-8533		128-bit LPDDR5x-7467
Size		96 GB		32 GB		128 GB
Bandwidth				136 GB/s
GPU		Intel Graphics	Intel Graphics	Arc 140V		Intel Graphics
RT		No	No	YES		YES
EU / Xe		96 EU	2 Xe	8 Xe		4 Xe
Max Clock		1.3 GHz	?	2 GHz		2.5 GHz
NPU		GNA 3.0	18 TOPS	48 TOPS		49 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.

 

[Saylick]

I'm not sure which profile was used for 14900K but according to CPU monkey, the SKU should score 38700 pts.
For reference 7950x gets 38.6K , while unofficially the 9950X gets 43.9K pts.

I think we’re realistically looking at ARL-S and the 9950X trading blows in MT, thanks to Chadmont. But the ST is anyone’s guess right now. What we need are ST clocks. Need to understand at what clocks are those ST scores being delivered at.

 

[cebri1]

I think we’re realistically looking at ARL-S and the 9950X trading blows in MT, thanks to Chadmont. But the ST is anyone’s guess right now. What we need are ST clocks. Need to understand at what clocks are those ST scores being delivered at.

This, without clocks the benchmark is mostly useless.

 

[inf64]

I think we’re realistically looking at ARL-S and the 9950X trading blows in MT, thanks to Chadmont. But the ST is anyone’s guess right now. What we need are ST clocks. Need to understand at what clocks are those ST scores being delivered at.

I think that the max boost clocks for ST will be 5.7 or 5.8Ghz, according to leaks. I guess that the QS sample was probably running at close to the max ST boost.

 

[csbin]

ARL-S QS

p
ST 5.7 MT 5.4

E
ST 4.7 MT 4.6

 

[SiliconFly]

-

 

[Philste]

250W PL2 is still not 250W power draw, even if we got used to it through last Intel gens. We don't know if this sample was able to run at final clocks, we don't know the actual Power draw.

 

[dullard]

I think the 14900K does a little better at CB23 than that Xitter post shows. For example, from TechPowerUp:...

One thing to keep in mind is that the data you posted are not at 250 W. Whereas the leak claims 250 W.

• TechPowerUp tested at ~280 W (hard to get an exact number from the top chart here: https://www.techpowerup.com/review/intel-core-i9-14900k/22.html) and got a Cinebench R23 multithread score of 38746.
• Anandtech tested at max 428 W (didn't breakout exact power in R23) and got a Cinebench R23 multithread score of 41211.
• Tweaktown tested at max 330 W (didn't breakout exact power in R23) and got a Cinebench R23 multithread score of 40525.

Also, what is the single thread power usage in the leak? Same ~35W to ~55W? Or different power levels between Arrow Lake and Raptor Lake Refresh?

 

[inf64]

ARL-S QS

p
ST 5.7 MT 5.4

E
ST 4.7 MT 4.6

It looks like the ARL-S ES part (first column in the X post) was running at 4.7Ghz while the QS sample ran at 5.7Ghz. The ST benchmarks align with it.

 

[jdubs03]

I'm not sure which profile was used for 14900K but according to CPU monkey, the SKU should score 38700 pts.
For reference 7950x gets 38.6K , while unofficially the 9950X gets 43.9K pts.

With PBO+EXPO enabled.

 

[inf64]

With PBO+EXPO enabled.

We are talking about intel CPUs here, not AMD. I was questioning the power profile of 14900K that was mentioned to score less than what we have in reviews.

 

[jdubs03]

We are talking about intel CPUs here, not AMD. I was questioning the power profile of 14900K that was mentioned to score less than what we have in reviews.

Indeed, just wanted to point that out the stock values as that it’s probably most representative of the QS from Intel. The percentage difference is a decent amount there.

 

[DavidC1]

nah, disregard my post, the issue is with the ST performance. Also MT is behind expectation, still a QS sample tho.

Actually.

If you take the assumption that 2 E cores are equal to 1 P cores for Raptorlake,

RPL: 8x1.3 + 8 = 18.4
ARL: 8x1.15 + 8 x 1.5 = 21.2(15%)

15% is roughly the figure for MT. We could say it's even outperforming expectations a bit as the above suggests MT clocks haven't went down from the 14900K.

6% in ST is basically 1.15 x 5.7/6.2.
(I know Intel said 14% but 1% is easily swayed by benchmark choices)

 

[Hitman928]

6% in ST is basically 1.15 x 5.7/6.2.
(I know Intel said 14% but 1% is easily swayed by benchmark choices)

Crossmark is a mix of workloads. The overall score will be influenced by the multithreaded performance. Looking at just the single/very lightly threaded results, it shows the ARL QS as being either the same or less performance than a 14900K (which is 6 GHz boost, not 6.2 GHz). It's not a good showing in terms of ST performance. Of course, it is just an unverified leak of a QS so there's nothing to conclude here, but if this is representative of retail performance, it's not going to be a good look for Intel.

Edit: Using GB5 ST results, the IPC increase would be 6.3%. The browser benchmarks show no IPC gain (WebXPRT) and a -4.5% regression (Speedometer). Leaked QS sample caveat obviously still applies.

 

[DavidC1]

but if this is representative of retail performance, it's not going to be a good look for Intel.

Yes, but we kinda knew the gains based on ARL leak a while ago that people dismissed as sandbagging. And the recent numbers are roughly on par with it.

The best we're going to get in single thread is 5-7%.

 

[Hitman928]

Yes, but we kinda knew the gains based on ARL leak a while ago that people dismissed as sandbagging. And the recent numbers are roughly on par with it.

The best we're going to get in single thread is 5-7%.

I added an edit, but these results don't even show that. It's basically a wash for ST performance at best. Of course, retail samples could perform better but there probably won't be a big difference.

 

[DavidC1]

I added an edit, but these results don't even show that. It's basically a wash for ST performance at best. Of course, retail samples could perform better but there probably won't be a big difference.

From Igorsleak: https://www.igorslab.de/en/intels-i...n-for-raptor-lake-s-refresh-and-arrow-lake-s/

Hmm, I guess it is.

From the Spec 1-copy results though, it only shows 2-3% gain over RPL-S Refresh which would be the 14900K.

 

[Hitman928]

From Igorsleak: https://www.igorslab.de/en/intels-i...n-for-raptor-lake-s-refresh-and-arrow-lake-s/

Hmm, I guess it is.

From the Spec 1-copy results though, it only shows 2-3% gain over RPL-S Refresh which would be the 14900K.

If this QS is retail performance and it is under performing compared to projections, they will probably miss SPEC 1T performance projection as well, so even that 2-3% gain is in question. Not drawing a conclusion yet, we are still a few months off from retail release, just commenting on the results we have.

 

[Bouowmx]

The separated core and memory tiles penalty rears its head in the browser benchmarks?

 

[Hitman928]

The separated core and memory tiles penalty rears its head in the browser benchmarks?

That could very well be the issue. The tile/chiplet penalty is hurting them more than they expected.

 

[DavidC1]

The separated core and memory tiles penalty rears its head in the browser benchmarks?

They would have known back when they were doing projections. Geekerwan's test put serious doubt on whether "tile penalty" is noticeable.

Still, there is a good reason to wait as few % can make a difference.

 

[Hitman928]

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.

 

[jdubs03]

Intel better hope they can achieve those values. Because it isn’t looking like it as of now.

 

[Philste]

Did you all miss the fact that it says "preliminary"? Jaykihn himself posted that Arrow Lake QS is Week 34, which is 4 weeks from now. I don't know where this data is from, but definitely not from a QS.

 

[Hitman928]

Did you all miss the fact that it says "preliminary"? Jaykihn himself posted that Arrow Lake QS is Week 34, which is 4 weeks from now. I don't know where this data is from, but definitely not from a QS.

Where did he post the date of the QS? It would be pretty strange to get a CPU to test that was manufactured in the future.

Other than that, everyone knows it's preliminary, that has been repeated multiple times now. If it really is a QS though, those are typically very close, if not equal, to retail samples. There's a reason they get listed as QS. That's not to say there can't be a difference, just that typically, there isn't a significant one between QS and retail, though it is possible.

 

[Det0x]

*edit*
Nevermind, already posted above 👍