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+4+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.

 

[misuspita]

It's just going to match it at the best.

AMD is just focusing on AI that's why. They could have gave a large cache if it wasn't for the Die Waste Unit.

That's exactly the point. AMD could have given us a much better product mix, but chose not to. Why, it doesn't matter. Still, I really hope Intel gets to parity so it forces AMD to give a better product mix at better prices. Halo is great, but too expensive. Compound it with RAM prices and it's even more unobtainium

 

[DavidC1]

Half-halo > PTL at same TDP easy

Yea at prices no one will buy.

Dude stop. Intel won this round. You like it when AMD wins one right? First time in a long time Intel wins.

 

[Hulk]

Yea at prices no one will buy.

Dude stop. Intel won this round. You like it when AMD wins one right? First time in a long time Intel wins.

I will be the first one cheering Intel's win. This would be an architectural and process (18A) node and would be fantastic.

But at this moment it is a "paper win."

Things are looking good but let's save the party for after the reviews.

 

[DavidC1]

I will be the first one cheering Intel's win. This would be an architectural and process (18A) node and would be fantastic.

But at this moment it is a "paper win."

Things are looking good but let's save the party for after the reviews.

I get your point, but this kind of an advantage(82% over Strix in native rendering) isn't something that can be significantly impacted by any such things. They are also claiming 50% at same W as Lunarlake.

I'm reminded of ATI R300. Previous to that, ATI would intro a competitive card and Nvidia would beat it with new drivers. I remember the Geforce 3 doing that with the Lightspeed memory architecture supporting driver, gaining 15% or so. When the R300 came out, the advantages were so big, no amount of driver changes would usurp it. As history shows, the follow up was NV30, which turned the script around, still lost to R300 that came a gen earlier, and cheated in image quality for "performance".

 

[MS_AT]

First time in a long time Intel wins.

Well, if product announcements could let you win something, then Intel never lost However the Panther Lake will do, I would wait for actual shelf products to measure its success. But thanks to Panther Lake our IT department might finally go AMD, they always seem to prefer the worse product lol.

 

[511]

Strix Point with 16CUs RDNA4 and 16-24MB Infinity Cache would probably give hell to Panther Lake, but unfortunately the lack of competition led AMD to sit back and relax. So now the only thing they can realistically use against Panther Lake is Strix Halo sold for cheap.

Big doubt for the cache it would be like increased gaming performance by double digit for sure but not >50% if it was RDNA 4 than yes it's possible to be in similar league as PTL or beat in games

 

[adroc_thurston]

Big doubt for the cache it would be like increased gaming performance by double digit for sure but not >50% if it was RDNA 4 than yes it's possible to be in similar league as PTL or beat in games

well yeah it's a smaller config.
12Xe is a 12WGP equivalent.

 

[Hulk]

I get your point, but this kind of an advantage(82% over Strix in native rendering) isn't something that can be significantly impacted by any such things. They are also claiming 50% at same W as Lunarlake.

I'm reminded of ATI R300. Previous to that, ATI would intro a competitive card and Nvidia would beat it with new drivers. I remember the Geforce 3 doing that with the Lightspeed memory architecture supporting driver, gaining 15% or so. When the R300 came out, the advantages were so big, no amount of driver changes would usurp it. As history shows, the follow up was NV30, which turned the script around, still lost to R300 that came a gen earlier, and cheated in image quality for "performance".

Your points are unassailable. I'm simply not giving Intel credit until the product performs as advertised in the hands of reviewers and consumers and is available in quantity. I don't trust Intel's announcements any more. Trust must be earned back.

 

[jpiniero]

Your points are unassailable. I'm simply not giving Intel credit until the product performs as advertised in the hands of reviewers and consumers and is available in quantity. I don't trust Intel's announcements any more. Trust must be earned back.

That's why I brought up the Frame Gen point. Better wait for reviews.

 

[DavidC1]

Your points are unassailable. I'm simply not giving Intel credit until the product performs as advertised in the hands of reviewers and consumers and is available in quantity. I don't trust Intel's announcements any more. Trust must be earned back.

There's a native rendering slide.

https://videocardz.com/newz/intel-says-arc-b390-graphics-are-73-faster-than-amd-radeon-890m-special-version-for-handheld-confirmed

I'm comparing with Notebookcheck results. It's fairly comparable. There are both results that are higher than NBC's results and lower than results. The absolute frames speaks for itself especially against Lunarlake. They were overall right about Lunarlake were they not?

Also, regarding the 4050 comparison. The 45W 4050 is 60-80% faster than the 25W Lunarlake(Thinkpad Gen 6) in average. Of course there are varying results depending on systems as well, since laptops are enormously sophisticated with varying power management and thermals. I get 20-25 fps on E33 with 1440p Med on a GTX 1080 for comparison.

The 25W Pantherlake should be competitive with 35W 4050 and 45W should be competitive with 45W. They said 60W but I'm looking at the conservative side. 25W and 45W doesn't seem like a huge difference(1.77 vs 1.5x which is only about 15%) except obviously on CPU bound titles where it needs more power for CPU to push frames. I'm actually surprised it can keep up in FPS titles like Valorant.

The challenge in Notebook comparisons are the following:
-In CPU bound titles and performance, the iGPUs are at a disadvantage, because the CPU can't flex fully. A system with same GPU but with higher end CPU will do better in that case. A 13th Gen HX with 35W GPU will be faster than 12th Gen H with 45W GPU.
-Thermal and power throttling, in both CPU and GPU
-Varying driver versions
-Sometimes, it just doesn't perform as well. Like you can see few % differences in desktop cards.
-Custom power management settings by the manufacturer. Look at the Intel Lunarlake guide about setting PL2 at least 1W higher than PL1.

 

[DZero]

Yea at prices no one will buy.

Dude stop. Intel won this round. You like it when AMD wins one right? First time in a long time Intel wins.

Funny story, seeing ARM... seems that Qualcomm and Apple won big time.

Decent evolution, but the improvements from ARM are notorious.

 

[mikk]

Your points are unassailable. I'm simply not giving Intel credit until the product performs as advertised in the hands of reviewers and consumers and is available in quantity. I don't trust Intel's announcements any more. Trust must be earned back.

You can see it on LNL/ARL reviews where the iGPU was already on par with AMDs dualchannel iGPU. PTL 12Xe has 50% more shaders, higher clock speed, twice the cache on a newer Xe3 architecture....this is a different league for dualchannel iGPUs. As for improved battery life it mainly comes from the LPE cores and fixed uncore, it can be as good as Lunar Lake. This was also expected.

 

[Doug S]

Things are looking good but let's save the party for after the reviews.

Yes, hot takes where people pronounce one side the winner or the other dead are all too common when we haven't seen actual benchmarks yet. Or worse hot takes based on speculation about how something in the future that isn't even announced with marketing benchclaims yet like Zen 6 or M6 or Intel 14A or whatever will do.

 

[Hulk]

There's a native rendering slide.

https://videocardz.com/newz/intel-says-arc-b390-graphics-are-73-faster-than-amd-radeon-890m-special-version-for-handheld-confirmed

I'm comparing with Notebookcheck results. It's fairly comparable. There are both results that are higher than NBC's results and lower than results. The absolute frames speaks for itself especially against Lunarlake. They were overall right about Lunarlake were they not?

Also, regarding the 4050 comparison. The 45W 4050 is 60-80% faster than the 25W Lunarlake(Thinkpad Gen 6) in average. Of course there are varying results depending on systems as well, since laptops are enormously sophisticated with varying power management and thermals. I get 20-25 fps on E33 with 1440p Med on a GTX 1080 for comparison.

The 25W Pantherlake should be competitive with 35W 4050 and 45W should be competitive with 45W. They said 60W but I'm looking at the conservative side. 25W and 45W doesn't seem like a huge difference(1.77 vs 1.5x which is only about 15%) except obviously on CPU bound titles where it needs more power for CPU to push frames. I'm actually surprised it can keep up in FPS titles like Valorant.

The challenge in Notebook comparisons are the following:
-In CPU bound titles and performance, the iGPUs are at a disadvantage, because the CPU can't flex fully. A system with same GPU but with higher end CPU will do better in that case. A 13th Gen HX with 35W GPU will be faster than 12th Gen H with 45W GPU.
-Thermal and power throttling, in both CPU and GPU
-Varying driver versions
-Sometimes, it just doesn't perform as well. Like you can see few % differences in desktop cards.
-Custom power management settings by the manufacturer. Look at the Intel Lunarlake guide about setting PL2 at least 1W higher than PL1.

All great new for Intel. Better efficiency than HX370 + better performance + iGPU on par with a 4050 is amazing. No doubt!

I will be Cheer Captain as soon as all of this is confirmed by reviewers and actual consumers!

 

[jdubs03]

Do we know when review embargoes lapse? I would imagine it’s before January 27 which I saw was the delivery dates for some devices.

 

[poke01]

IMO, the GPU is impressive because AMD is behind in IP and node.

If we actually take a look at a current iGPU IP on a current node on newer titles that are based around UE5 or newer engines based on ray tracing, Intel's GPU IP doesn't appear that far ahead. They just brute force the GPU with more cache and titles like Cyberpunk 2077 like it.





If we compare Assassin Creed shadows which is native on Mac at 1080p High, the M5 GPU does 26 fps and the B390 does 30 fps. Yes, the PTL GPU is faster by around 15% but it also has 5 TIMES the L2 cache (16mb) for the GPU whereas the M5 GPU has around ~3MB of L2 cache.

Assassin's Creed Shadows review: High-fidelity visuals, next-gen textures and brutal demands on your hardware

We put Assassin's Creed Shadows through a rigorous technical and benchmark test across a range of laptop and desktop GPUs.

www.notebookcheck.net

 

[Joe NYC]

IMO, the GPU is impressive because AMD is behind in IP and node.

Exactly. And RDNA3 is not AMD's most fortunate generation.

Then, skipping RDNA4, making the gap from last decent generation extra long (RDNA2 -> RDNA5).

So, 2026 will be the lowest point for AMD in iGPU.

 

[mvprod123]

IMO, the GPU is impressive because AMD is behind in IP and node.

If we actually take a look at a current iGPU IP on a current node on newer titles that are based around UE5 or newer engines based on ray tracing, Intel's GPU IP doesn't appear that far ahead. They just brute force the GPU with more cache and titles like Cyberpunk 2077 like it.

View attachment 136323


View attachment 136325
If we compare Assassin Creed shadows which is native on Mac at 1080p High, the M5 GPU does 26 fps and the B390 does 30 fps. Yes, the PTL GPU is faster by around 15% but it also has 5 TIMES the cache (16mb) for the GPU whereas the M5 GPU has around ~3MB of L2 cache.

Assassin's Creed Shadows review: High-fidelity visuals, next-gen textures and brutal demands on your hardware

We put Assassin's Creed Shadows through a rigorous technical and benchmark test across a range of laptop and desktop GPUs.

www.notebookcheck.net

This is considering that none of the native Mac games have been updated to Metal 4 yet.

 

[poke01]

This is considering that none of the native Mac games have been updated to Metal 4 yet.

true however native AAA games on Mac are just a technical showcase for new Apple hardware IP. game devs just don't give a hoot for the latest APIs on the mac.

The takeaway is that if Intel stripped half of its l2 cache away would we still see major GPU improvements, I don't think so.
This is the most expensive iGPU for a 128-bit SOC period and it better be strong for much cache it packs.

 

[DavidC1]

true however native AAA games on Mac are just a technical showcase for new Apple hardware IP. game devs just don't give a hoot for the latest APIs on the mac.

The takeaway is that if Intel stripped half of its l2 cache away would we still see major GPU improvements, I don't think so.
This is the most expensive iGPU for a 128-bit SOC period and it better be strong for much cache it packs.

With Cyberpunk 2077, they claimed 27% reduction in access to memory, which equals to 30% difference in bandwidth. That's only about 10% difference in performance. In average, the performance is split roughly equally between pixel pushing, bandwidth, and shader performance.

The average reduction in traffic is about 20%, which is even less. The advantages is a combination of more shaders, cache, software, power management, and uarch. The older Intel uarchs including Xe2 suffers from shader utilization issue as well, which improves in Xe3 with dynamic register allocation. There's also the improvement for occlusion culling which will improve small data size performance such as instructions versus larger caches which doesn't. Just increasing shaders by 50% would have resulted in 25-35%.

Also, SRAM is cheap in both power and cost wise. It saves power by preventing accesses out of the memory, and cost wise it's extremely redundant. It's the most copy-paste use of transistors. Since M5 GPU is monolithic which doesn't waste with redundancies, and it's already 33mm2 with M4, I doubt it's noticeably cheaper either.

Yes, hot takes where people pronounce one side the winner or the other dead are all too common when we haven't seen actual benchmarks yet. Or worse hot takes based on speculation about how something in the future that isn't even announced with marketing benchclaims yet like Zen 6 or M6 or Intel 14A or whatever will do.

The only way Pantherlake will not beat it is if they straight up lie on their numbers. Most complaints are that the competing system is using disadvantaged setups, which isn't the case here. When people were talking about Arrowlake vs Zen 5, or 890M vs Lunarlake, the differences were close enough that it wasn't a decisive win until post release. It's not the case here.

I've seen the process+uarch arguments before as well. It was hoped(including me) that core count parity and new process would make Intel competitive with Granite Rapids. That wasn't the case. For the core count, Emerald Rapids were very well done, and we did scaling based on that.

 

[regen1]

Core Series 3(Wildcat Lake) is mentioned at around 12 minute mark.

 

[DavidC1]

Core Series 3(Wildcat Lake) is mentioned at around 12 minute mark.

70mm+40mm2 used to be called expensive. Not anymore apparently. If somehow they got 450mm wafers working, they would have gotten a long term benefit. That's why computer prices have been reversing and increasing over time.

 

[poke01]

Also, SRAM is cheap in both power and cost wise. It saves power by preventing accesses out of the memory, and cost wise it's extremely redundant. It's the most copy-paste use of transistors. Since M5 GPU is monolithic which doesn't waste with redundancies, and it's already 33mm2 with M4, I doubt it's noticeably cheaper either.

The B390 with L2 is 53-54mm2, the M5 GPU with L2 is ~35-36mm2. The B390 is almost 50% larger. So PPA could be improved.

That being said Intel has gotten a major win and with AMD releasing RDNA 3.5 again, its an easy win for Intel.

 

[branch_suggestion]

The B390 with L2 is 53-54mm2, the M5 GPU with L2 is ~35-36mm2. The B390 is almost 50% larger. So PPA could be improved.

That being said Intel has gotten a major win and with AMD releasing RDNA 3.5 again, its an easy win for Intel.

Well AMD canned the RDNA4.5 roadmap that would've dealt with it.
It is a nice niche, won't last forever but congrats Intel for finding a free opening.

 

[DZero]

IMO, the GPU is impressive because AMD is behind in IP and node.

If we actually take a look at a current iGPU IP on a current node on newer titles that are based around UE5 or newer engines based on ray tracing, Intel's GPU IP doesn't appear that far ahead. They just brute force the GPU with more cache and titles like Cyberpunk 2077 like it.

View attachment 136323


View attachment 136325
If we compare Assassin Creed shadows which is native on Mac at 1080p High, the M5 GPU does 26 fps and the B390 does 30 fps. Yes, the PTL GPU is faster by around 15% but it also has 5 TIMES the L2 cache (16mb) for the GPU whereas the M5 GPU has around ~3MB of L2 cache.

Assassin's Creed Shadows review: High-fidelity visuals, next-gen textures and brutal demands on your hardware

We put Assassin's Creed Shadows through a rigorous technical and benchmark test across a range of laptop and desktop GPUs.

www.notebookcheck.net

And the competitor is supposed to be the M5 Pro... not the M5 due costs.