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]

Well the problem is that if Arrowlake being a 2024 product with 8P cores means that Intel does not have a good answer against Zen 4 3D V-Cache next year. Like I don't see the 13900KS being a good solution against V-Cache CPUs who'll be power sipping in comparison.

I don't think Intel is focusing on gaming crown with MTL. They need to fix mobile first mainly the ultrabook and high end laptop market which is far more important than having a fast desktop gaming CPU.

It's rumoured that Arrow Lake will launch closely to MTL and that is more focused on desktop and mobile.

 

[poke01]

https://twitter.com/x/status/1529763542978936832

Remember Lunar Lake is going to be mobile only and for 15watts under maybe.

 

[DrMrLordX]

It's cuz they clocked the e-cores so high

That's part of the problem, yes. There's also (arguably) too many of them.

 

[IntelUser2000]

I don't think Intel is focusing on gaming crown with MTL. They need to fix mobile first mainly the ultrabook and high end laptop market which is far more important than having a fast desktop gaming CPU.

Desktop is important, just less important. They said the Desktop volume went up 2% due to the strength of gaming and enthusiast sales. Also, enthusiasts have a greater effect on the market due to their knowledge guiding people around them. Intel research said 1 enthusiast affects the decision making process of average a dozen people or something.

That's why Arrowlake is coming with Meteorlake, because especially in a weak market, they cannot neglect anything. Of course mobile is important too. Meteorlake with the ULP E cores(dedicated to shutting down blocks) seem like a much better mobile product for sure. I think even Raptorlake might be a good step. Well, that might be a strong word since Alderlake just has lots of problems on the battery life side and potentially Raptorlake would just fix that issue.

 

[poke01]

They said the Desktop volume went up 2% due to the strength of gaming and enthusiast sales

DIY is a drop in an ocean. Most of the money and profit is made from mobile and datacentre. Yes, desktop is important but it takes a backseat to mobile and datacentre.

 

[IntelUser2000]

DIY is a drop in an ocean. Most of the money and profit is made from mobile and datacentre. Yes, desktop is important but it takes a backseat to mobile and datacentre.

Not anymore according to the earnings sheet. Notebook share of revenue went from 65.6% to 57.8% in a single year. Entire DCAI revenue is just 31% larger than Desktop revenue now. Note that I've entirely excluded the Other section of client which includes things like WiFi modems.

Desktops are more resilient and the groups in Intel learn more because the enthusiasts are a picky bunch. We give them hell because we have to deal with individual upgradeable components rather than a black box aka notebooks. I learned electronics thanks to the DIY market. I threw away a motherboard when I was a kid because I flashed it wrong. I now can reprogram it with a proper one just by buying a new firmware chip.

Percentage-wise it's less but each percent has greater value because of that. That's why the PC market is still strong despite the 2-decade claims of "PC is dead".

 

[controlflow]

Well the problem is that if Arrowlake being a 2024 product with 8P cores means that Intel does not have a good answer against Zen 4 3D V-Cache next year. Like I don't see the 13900KS being a good solution against V-Cache CPUs who'll be power sipping in comparison.

Is Zen 4 actually more efficient than RPL in gaming? Der8auer showed that the 13900k pretty much maintains the same gaming performance at 90w vs full power.

A lot of people seem to assume that Zen 4 3D will be to Zen 4 what Zen 3 3D was for Zen 3. I'm not sure we can make that assumption. Both Zen 4 and RPL have already improved the cache deficiency situation from their predecessors and you may run into a case of diminishing returns with added cache on Zen 4 3D. It may be a big improvement but we can't just blindly assume that. Adding cache doesn't just magically make all games run faster. Once the cache is no longer the limiting factor, you no longer get improvement.

We have plenty of benchmarks with a 13900k and a 4090 showing many cases where it is leading AMD parts by 10-15%+ and even more in 1% minimums. IMO it may be that Zen 4 3D may not be much better than a future 13900KS and DDR5 7000+ if it is faster at all.

 

[Markfw]

Is Zen 4 actually more efficient than RPL in gaming? Der8auer showed that the 13900k pretty much maintains the same gaming performance at 90w vs full power.

A lot of people seem to assume that Zen 4 3D will be to Zen 4 what Zen 3 3D was for Zen 3. I'm not sure we can make that assumption. Both Zen 4 and RPL have already improved the cache deficiency situation from their predecessors and you may run into a case of diminishing returns with added cache on Zen 4 3D. It may be a big improvement but we can't just blindly assume that. Adding cache doesn't just magically make all games run faster. Once the cache is no longer the limiting factor, you no longer get improvement.

We have plenty of benchmarks with a 13900k and a 4090 showing many cases where it is leading AMD parts by 10-15%+ and even more in 1% minimums. IMO it may be that Zen 4 3D may not be much better than a future 13900KS and DDR5 7000+ if it is faster at all.

I can say this. Zen 4 at default is more efficient than Raptor lake, but in gaming it may or may not be. However, setting it to ECO mode (and the same for Raptor Lake) I am sure Zen 4 can be more efficient. Somebody needs to test this scenario to be sure.

 

[DrMrLordX]

However, setting it to ECO mode (and the same for Raptor Lake)

Does Intel actually have an equivalent to ECO mode? You can certainly limit PL values in the UEFI, but that's quite a bit different than setting a software switch to ECO1 or ECO2, rebooting, and being good to go (or doing the same in a UEFI menu).

 

[Markfw]

Does Intel actually have an equivalent to ECO mode? You can certainly limit PL values in the UEFI, but that's quite a bit different than setting a software switch to ECO1 or ECO2, rebooting, and being good to go (or doing the same in a UEFI menu).

Actually, my setup does not actually involve ECO mode. I set the PBO CO option to -25.

 

[DrMrLordX]

Actually, my setup does not actually involve ECO mode. I set the PBO CO option to -25.

Right, but your average plug n' play type of user doesn't necessary want to fiddle with all of that. AMD has a software switch in Ryzen Master that lets you toggle 105W TDP and 65W TDP ECO modes. I don't think Intel XTU has the same thing, though having not used it, I'm not 100% sure about that.

 

[Markfw]

Right, but your average plug n' play type of user doesn't necessary want to fiddle with all of that. AMD has a software switch in Ryzen Master that lets you toggle 105W TDP and 65W TDP ECO modes. I don't think Intel XTU has the same thing, though having not used it, I'm not 100% sure about that.

Trying to be fair to Intel. Its not as easy, but set pl1 and pl2 to 150 watt or so, and its the same as what I did. Except I can outperform them easily if we are both at the same power.

 

[eek2121]

MTL will be great in mobile. Arrow Lake will be excellent in mobile. I am not excited for desktop anymore. It's more fun to see what Intel/AMD do in thermal constrained devices.

As efficiency becomes more important mobile will reap the biggest benefit this decade.

It would actually be great if we could get standardized form factors for laptops and mini PCs. Combine that with DIY replaceable parts. Some companies have tried, sure, but you end up suffering from vendor lock-ins. If standards bodies did it and hardware companies supported it, I bet we would see a shift in DIY.

I bring this up because traditionally many have gone with gaming desktops because they have no other choice. Laptops and mini PCs typically have too many compromises.

DIY with mid/large towers is a replica from the 80s. It doesn’t have to be that way.

If you could buy a motherboard, RAM, and GPU of choice to combine with your laptop keyboard, chassis, screen, and battery of choice, that would be amazing.

Why on earth would you think that something with 30% to 50% more cores (albeit E cores), a tweaked core design, almost 10% faster turbo, 17% faster memory, and more cache would be just 3% to 5% faster? Heck the 5% higher turbo power level alone would let Raptor Lake be more than 3% faster than Alder Lake even if you ignored all the other changes.

I've always been in the camp that Raptor Lake would be much better than Alder Lake but Meteor Lake is going to disappoint many people (wait for Arrow Lake). Unless someone selectively picks only bad rumors and not any of the good rumors, then there had to be a much more significant speed bump than "at tops 3% to 5%".

Always take pre-release benchmarks with a large grain of salt. Your 3% to 5% number can be changed with a BIOS update alone.

I think you are confusing numbers. I don’t think anyone claimed 3-5% for Raptor Lake. The 3-5% was for IPC increase of the golden cove cores relative to alder lake. Everyone knew the other changes would bring an uplift, and many of us knew Intel would have to raise power limits to do it. Raising power limits kind of defeats the purpose. Raptor Lake is on of the most inefficient x86 CPUs out there currently.

Performance in my eyes was never in question. Efficiency was. Intel would not be competing with Zen 4 if they stuck to a ~230W PL2 (which the chip was originally specced at according to leaks 2 years ago, note that is slightly lower than Alder Lake)

We all know Intel would brute force things with power. It is a bad habit for them. It is something that keeps me from buying their products (that and the constant socket changes)

It has honestly become a bad habit for the entire industry.

I am hoping Meteor Lake and Arrow Lake changes this. I suspect it won’t.

 

[igor_kavinski]

I am hoping Meteor Lake and Arrow Lake changes this. I suspect it won’t.

Transistor density of Intel 4 would prevent Intel from approaching 300W. The heat would be enough to burn out the CPU.

 

[Doug S]

If you could buy a motherboard, RAM, and GPU of choice to combine with your laptop keyboard, chassis, screen, and battery of choice, that would be amazing.

The size of that market would be minuscule, that's the reason it doesn't exist. The DIY market for desktops used to be (long ago, I'm talking in the 90s) about value - you could get more for your money if you bought the parts and built the system yourself even if you just copied specs available from HP et al. By the early/mid 2000s that was no longer true, and today you can't come close to building the same stuff via parts except in certain situations (i.e. the price of components are falling quickly like we're seeing right now with SSDs and DRAM and the OEMs aren't cutting prices / upping specs at the same pace)

So now the DIY market is basically gamers and a few weirdos like me who build systems because I want to make it as silent as possible.

It is far more likely the DIY market eventually disappears for desktops than a DIY market appearing for laptops.

 

[moinmoin]

The size of that market would be minuscule, that's the reason it doesn't exist. The DIY market for desktops used to be (long ago, I'm talking in the 90s) about value - you could get more for your money if you bought the parts and built the system yourself even if you just copied specs available from HP et al. By the early/mid 2000s that was no longer true, and today you can't come close to building the same stuff via parts except in certain situations (i.e. the price of components are falling quickly like we're seeing right now with SSDs and DRAM and the OEMs aren't cutting prices / upping specs at the same pace)

So now the DIY market is basically gamers and a few weirdos like me who build systems because I want to make it as silent as possible.

It is far more likely the DIY market eventually disappears for desktops than a DIY market appearing for laptops.

I don't agree about the disappearing part. Aside value I'd say DIY is mainly about choice and quality. With laptops and OEM desktops you get a pig in a poke, and if you have very specific but uncommon requirements there likely is no offer on the market that can match it. Unless you or somebody else custom-builds it for you which is what DIY enables and makes it valuable beyond "gamers and a few weirdos".

 

[IntelUser2000]

The DIY market for desktops used to be (long ago, I'm talking in the 90s) about value - you could get more for your money if you bought the parts and built the system yourself even if you just copied specs available from HP et al.

It's still true today but you need to take a long view.

Storage, Case, Power Supply, Monitor, Keyboard, Mouse doesn't need to be replaced. Storage upgrades are overrated, even the SATA II ones are still very good. Maybe you can upgrade to PCI Express 4.0 one but again that'll last you for 5 years and only if sequential upgrades matter to you.

So in the beginning it costs more, but later all you change is CPU+Motherboard, and GPU. RAM lasts for many years as well. Try competing with that!

 

[IntelUser2000]

Performance in my eyes was never in question. Efficiency was. Intel would not be competing with Zen 4 if they stuck to a ~230W PL2 (which the chip was originally specced at according to leaks 2 years ago, note that is slightly lower than Alder Lake)

Raptorlake does better in perf/watt relative to Zen 4 than Alderlake did compared to Zen 3. And it is a great improvement over Raptorlake.

It has honestly become a bad habit for the entire industry.

I am hoping Meteor Lake and Arrow Lake changes this. I suspect it won’t.

I don't know what you mean by "change"? Surely it will change. Raptorlake improved the competitive position. Meteorlake(and Arrowlake further) is expected to improve on it more.

I assure you if Arrowlake is a great perf/watt gain, then you'll see 14900K, 20A, Arrowlake chip being set at 300W so it performs greatly better. You can always get a T CPU if you don't want to bother with the power settings. And very likely if you are posting in this forum, you can take 1-2 minutes to lower the power settings.

Industry is just following user demand. There are videos out there that uncork the limit of low power Gemini Lake platforms so it can use 15W rather than the default 6W for mobile and 10W for desktop, because it improves performance.

When you read reviews(for example from Notebookcheck), they criticize a 12W PL1 chip for not having enough performance. Well DUH!! But they don't see the reason behind the manufacturer setting it that low, and want to set it higher. For them, a 28W marketed chip like Alderlake-P has to be set at 28W or higher. So manufacturers generally set it at that level, or even higher to perform better. You can see PL1 settings of 35W-40W on some "28W" systems!

 

[Markfw]

Raptorlake does better in perf/watt relative to Zen 4 than Alderlake did compared to Zen 3. And it is a great improvement over Raptorlake.



I don't know what you mean by "change"? Surely it will change. Raptorlake improved the competitive position. Meteorlake(and Arrowlake further) is expected to improve on it more.

I assure you if Arrowlake is a great perf/watt gain, then you'll see 14900K, 20A, Arrowlake chip being set at 300W so it performs greatly better. You can always get a T CPU if you don't want to bother with the power settings. And very likely if you are posting in this forum, you can take 1-2 minutes to lower the power settings.

Industry is just following user demand. There are videos out there that uncork the limit of low power Gemini Lake platforms so it can use 15W rather than the default 6W for mobile and 10W for desktop, because it improves performance.

When you read reviews(for example from Notebookcheck), they criticize a 12W PL1 chip for not having enough performance. Well DUH!! But they don't see the reason behind the manufacturer setting it that low, and want to set it higher. For them, a 28W marketed chip like Alderlake-P has to be set at 28W or higher. So manufacturers generally set it at that level, or even higher to perform better. You can see PL1 settings of 35W-40W on some "28W" systems!

I would possibly argue that. My 7950s's are set to PBO CO -25, and use 142 watt, and they are killing my Alderlake (core for core 8 P-cores to 8 Zen cores), and Raptorlake does not even have 512, and I use it almost daily on DC projects. So my argument would be that perf/watt is highly dependent on the task, and if you altered the PL1 and PL2 to a particular value. In general, Zen 4 to Raptorlake can be better than Zen 3 to Alderlake.

 

[poke01]

The DIY market is gamers, enthusiasts and money/value oriented people. I am console gamer at heart but I am an hardware enthusiast. I look forward to the advancements from Intel and AMD. For me personally, Intel for laptops advancements(That's why I want Arrow/MTL/Lunar to be industry leading) and AMD for gaming/console/graphics improvements.

Each segment has its own value. People still use Nokia-style keypad phones these days and those are still being sold brand new.

 

[poke01]

When you read reviews(for example from Notebookcheck), they criticize a 12W PL1 chip for not having enough performance. Well DUH!! But they don't see the reason behind the manufacturer setting it that low, and want to set it higher. For them, a 28W marketed chip like Alderlake-P has to be set at 28W or higher. So manufacturers generally set it at that level, or even higher to perform better. You can see PL1 settings of 35W-40W on some "28W" systems!

I blame Intel. Intel needs to have a simple line up for ultrabooks. When Lunar Lake comes Intel first needs to tout it's pref/w advancements. Like force OEMs that use Lunar Lake to use up to 15 watt. This ensures that Intel chips have great battery life and pref. If they do this mainstream users and reviewers will claim Intel is efficient again. It's rumoured that Lunar is the "little guy". Also if Lunar uses Lion Cove it should still be very powerful chip at 15 watts.

No, 99.9% of normal users do not know what PL1/PL2 is that's why it's up to Intel ensure Lunar will not exceed 15watts even when boosting.

 

[IntelUser2000]

Like force OEMs that use Lunar Lake to use up to 15 watt. This ensures that Intel chips have great battery life and pref. If they do this mainstream users and reviewers will claim Intel is efficient again. It's rumoured that Lunar is the "little guy". Also if Lunar uses Lion Cove it should still be very powerful chip at 15 watts.

Nah, this is not a disadvantage, because it allows for design flexibility. If it was fixed, then Intel would have to offer multiple different chips just differing in TDP settings. Or you'd have one laptop design trying to serve all segments.

You can't fault them for the reviewers being retarded. Perhaps more education for the press?

One thing they can do is encourage manufacturers to be more open in what power settings their laptop uses. So Performance = 30W, Normal = 18W, Power Saving = 12W.

 

[poke01]

Nah, this is not a disadvantage, because it allows for design flexibility. If it was fixed, then Intel would have to offer multiple different chips just differing in TDP settings. Or you'd have one laptop design trying to serve all segments.

You can't fault them for the reviewers being retarded. Perhaps more education for the press?

One thing they can do is encourage manufacturers to be more open in what power settings their laptop uses. So Performance = 30W, Normal = 18W, Power Saving = 12W.

I believe Intel needs their M1 moment. Just focus on ultrabooks with Lunar Lake and cap the chip to 15watts. You still have Arrow Lake for beefier laptops/desktops.

For once I want an M1 like x86 chip focused on maximum efficiency that also delivers great pref for its class. A Surface Pro or an XPS 13" would be so good with Lunar Lake inside of it. All Intel needs to prove is that it can deliver the same battery life as Apple.
This would kill any Nuvia/ARM creep over in the laptop industry. People mainly want ARM because of battery life.

ARM chips don't do crazy boosting that go above their rated TDP that's part of the reason why ARM is cool and has lots of battery life.

 

[IntelUser2000]

This would kill any Nuvia/ARM creep over in the laptop industry. People mainly want ARM because of battery life.

The difference in TDP is not that big. Battery life is not directly correlated to TDP. The ARM platform, and even AMD has an advantage over Intel(especially Alderlake) in power management and idle.

Also if you had a higher TDP Apple Mx chip that performed higher than people would get that instead. Performance takes priority.

Future Intel generations will improve this. Alderlake regressed over the predecessor and it wouldn't be a surprise if Raptorlake fixes this. Meteorlake has ULP E cores dedicated to shutting down blocks and has a chance to greatly improve things.

 

[A///]

Does Intel actually have an equivalent to ECO mode? You can certainly limit PL values in the UEFI, but that's quite a bit different than setting a software switch to ECO1 or ECO2, rebooting, and being good to go (or doing the same in a UEFI menu).

through a software interface? there's xtu but I'm not sure what it's like on Intel. AMD's ryzen master and adrenalin look better than their contemporary counterparts software suites. Intel could do the same if they wanted to the eky word being wanted to. there's no incentive for them to waste man hours making things prety.