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.

 

[Hulk]

In terms of performance the 265k has been equal or better since I limited the 14900k to 200w. (Prefer reasonably quiet heatpipe based cooling over liquid.) I'm sure that raises an equally valid question of why I bothered with the 14900k in the first place. Answer there is I've been upgrading every generation and re-selling the used CPU+motherboard on eBay for roughly the original price paid since core 2 duo. Perk of having started my career at Intel and still having friends who work there is access to their employee purchase program. The at-cost pricing is far, far better than the 15-20% reimbursement AMD offers, or the 10% discount with NVIDIA. I had thought about skipping ARL, but the abnormally cheap pricing of the 265k combined with higher end z890 boards being re-sold at 1/2 to 1/3 of original price earlier this year made it a bit too tempting. (Thanks Newegg for your forced bundling of motherboards no one wanted with graphics cards far too many people wanted.)

I'd certainly like it if there were reasonable 2x performance opportunities to be had. It was fun getting those kinds of improvements from graphics cards for a bit without having to pay an obscene amount.

I was on the Raptor Lake train for a few years. They were good cpu's if run at sensible frequency/power. But at sensible frequency/power the 9950x runs circles around them. At least for my use cases it does. I'm excited to see what Nova brings to the table but for me to move to a new socket it would have to literally KO Zen 6.

 

[regen1]

Announcing hardware-accelerated BitLocker - Windows IT Pro Blog

BitLocker takes advantage of the latest and modern SoC and CPU capabilities for better performance and security.

techcommunity.microsoft.com

We have enhanced the architecture and implementation of the Windows storage and security stacks to support these new capabilities as an operating system enhancement that will bring value to all capable PCs over time. Upcoming Intel vPro® devices featuring Intel® Core™ Ultra Series 3 (formally codenamed Panther Lake) processors will provide initial support for these capabilities with support for other vendors and platforms planned.

 

[LightningZ71]

Impressive that they are finally moving to, or at least enabling, such a solution on the PC side. This has been a Mac feature for some time to my understanding with the original T1 chip to it's integration onto their current SoCs directly.

 

[511]

Impressive that they are finally moving to, or at least enabling, such a solution on the PC side. This has been a Mac feature for some time to my understanding with the original T1 chip to it's integration onto their current SoCs directly.

Yeah should have done sooner with windows 11 as a baseline

 

[Joe NYC]

In terms of performance the 265k has been equal or better since I limited the 14900k to 200w. (Prefer reasonably quiet heatpipe based cooling over liquid.) I'm sure that raises an equally valid question of why I bothered with the 14900k

Not as much why you got 14900k, buy mainly, if you already have one, and it is not showing signs of degradation, why move to 265k. The change in performance is a rounding error, for a lot of effort.

in the first place. Answer there is I've been upgrading every generation and re-selling the used CPU+motherboard on eBay for roughly the original price paid since core 2 duo. Perk of having started my career at Intel and still having friends who work there is access to their employee purchase program. The at-cost pricing is far, far better than the 15-20% reimbursement AMD offers, or the 10% discount with NVIDIA.

So, you are paying a lower price than what can be fetched by selling on E-Bay - that's a good incentive. But you still need to pay for mobo, swap mobo, and the mobo you get with Arrow Lake is already obsolete, with no upgrade path. So why not just sell the brand new 265k on Ebay (pocket the difference) and keep what you have?

I'd certainly like it if there were reasonable 2x performance opportunities to be had. It was fun getting those kinds of improvements from graphics cards for a bit without having to pay an obscene amount.

You will never get 2x if you upgrade so frequently. On my 2 systems, I went from:
work: Intel 2400lk -> AMD 5800x
home: Intel 5930k (HEDT) -> AMD 7800x3d

I think I did get roughly 2x in the blended ST / MT performance, weighted more to ST.

I could bend the 2x rule if it is CPU only upgrade, within AM5 motherboard (to Zen 6 or Zen 7).

On GPU side, I don't see myself upgrading for another 1-2 generations, which should be easy 2x performance increase (from 7800x), and still stay price-wise within the mainstream GPU segment of $500-$600 range.

 

[Joe NYC]

I'm excited to see what Nova brings to the table but for me to move to a new socket it would have to literally KO Zen 6.

BTW, that's half of the reason why AMD is outselling Intel in DIY by wide margin. Even with your 2nd gen of AM5 CPU, you will still be able to do a CPU only upgrade for next 2 full generations of CPUs.

Intel is completely missing out on CPU only upgrades.

Even if NVL matches Zen 6. For AM5 socket owners, Zen 6 will be a much more compelling upgrade. And judging by the rates of sales, by the time NVL is released, majority of DIY crowd may already be on AM5.

I really have no idea what Intel management was thinking, releasing 1 generation only socket design for Arrow Lake. Burning bridges with motherboard makers.

But at least (the rumor has it) that NVL socket may be more than 1 generation.

 

[511]

BTW, that's half of the reason why AMD is outselling Intel in DIY by wide margin. Even with your 2nd gen of AM5 CPU, you will still be able to do a CPU only upgrade for next 2 full generations of CPUs.

Zen 7 is not guaranteed to be AM5

Intel is completely missing out on CPU only upgrades.

Even if NVL matches Zen 6. For AM5 socket owners, Zen 6 will be a much more compelling upgrade. And judging by the rates of sales, by the time NVL is released, majority of DIY crowd may already be on AM5.

What about new buyers and I can guarantee DIY People sometimes make bad decision like overpaying for X3D I wouldn't pay that much unless that X3D is making money for me I would rather but non X3D than to pay the premium.

I really have no idea what Intel management was thinking, releasing 1 generation only socket design for Arrow Lake. Burning bridges with motherboard makers.

It was designed for 3 generation but we got 1 cause the 2 were canned

But at least (the rumor has it) that NVL socket may be more than 1 generation.

 

[Joe NYC]

Zen 7 is not guaranteed to be AM5

I think that by the time frame of Zen 7, or around 2028, we may see a proliferation of APUs like Strix/Medusa Halo and Intel Serpent Lake that will not even have a socket.

If that happens, then AMD will be happy to continue with AM5 on DDR5 (which already also has PCIe 5), and then also make an alternate push for socket-less Mini PC designs.

As long as Intel does not put pressure on AMD with some sort of HEDT design which could have more than 2 memory channels of DDR5, I don't see AMD releasing incompatible DDR5 platform with extremely limited upside to it.

With NVL, Intel will have 3 different DDR5 platform. People would be hard pressed to name the compelling upsides from 1st to 2nd and 3rd.

What about new buyers and I can guarantee DIY People sometimes make bad decision like overpaying for X3D I wouldn't pay that much unless that X3D is making money for me I would rather but non X3D than to pay the premium.

I am not CPU limited with anything work / pleasure / infotainment related outside of gaming.

So, it makes sense to address the only area where I am CPU bottlenecked with the best tool to relieve that bottleneck - V-Cache.

V-Cache also masks the memory latency. Subsequent (potential) CPU upgrades will likewise mask the memory latency, and the gobs of cheap memory I got a while back will do just fine. So, another reason V-Cache is paying for itself.

While you can't turn around on a dime, Intel still took a long time to address the challenge V-Cache presented. The brute force solution that Intel is promising for NVL should have been in place for ARL.

It was designed for 3 generation but we got 1 cause the 2 were canned

Still, can you name the compelling enhancement in 3 separate DDR5 sockets that are so compelling as to justify breaking compatibility?

 

[Khato]

Not as much why you got 14900k, buy mainly, if you already have one, and it is not showing signs of degradation, why move to 265k. The change in performance is a rounding error, for a lot of effort.

You say effort, I say fun! No, really, I've enjoyed tinkering with computer hardware since building my first around a K6-2. Started enjoying it even more once windows evolved to the point where it gracefully adapts to hardware changes instead of requiring a clean install.

As for the performance difference... High quality AV1 transcoding is the only task I run on this computer which is CPU bound. I'd forgotten that I have some performance numbers for the 14900k when I was testing 4k encoding settings, so I'll see how the 265k does running the same job by comparison.

So, you are paying a lower price than what can be fetched by selling on E-Bay - that's a good incentive. But you still need to pay for mobo, swap mobo, and the mobo you get with Arrow Lake is already obsolete, with no upgrade path. So why not just sell the brand new 265k on Ebay (pocket the difference) and keep what you have?

As noted, the motherboard was obtained for under 1/2 normal price. And the motherboard isn't 'obsolete' when it gets sold with the processor whenever I next upgrade. If anything that approach makes the 'upgrade path' easier than being locked into a specific socket because it's 'upgradable'. I could provide actual examples, but the short summary is that even without the CPU discounts it really isn't very expensive at all to upgrade CPU + motherboard when selling the used hardware. That fact rarely seems to be mentioned in the socket compatible upgrade path arguments. (Aside, in addition to the occasional good deal on eBay due to forced bundling, the other great source for motherboard deals would be used ones via Amazon Resale, so long as you accept that you're basically playing the part of hardware testing services in exchange for the price.)

 

[511]

Still, can you name the compelling enhancement in 3 separate DDR5 sockets that are so compelling as to justify breaking compatibility?

With NVL, Intel will have 3 different DDR5 platform. People would be hard pressed to name the compelling upsides from 1st to 2nd and 3rd.

More pins can be used for other purpose more power it's not like LGA 1150/51 where we got like 1 more pins and force socket change more features an AM5 X series board has less IO Features than a B860 Board ....

The reason is power there was 8+32 Planned for ARL also 16+32 is planned for NVL you need power delivery other stuff for such high core count also more PCI-E Lanes and connectivity Integrated TB5 for NVL do there are reasons for newer platform.

 

[Thibsie]

More pins can be used for other purpose more power it's not like LGA 1150/51 where we got like 1 more pins and force socket change more features an AM5 X series board has less IO Features than a B860 Board ....

The reason is power there was 8+32 Planned for ARL also 16+32 is planned for NVL you need power delivery other stuff for such high core count also more PCI-E Lanes and connectivity Integrated TB5 for NVL do there are reasons for newer platform.

Yeah let’s try to justify Intel’s ‘you will change your mobo every year’ by ´Intel is such incompetent that they didn’t thought maybe it could plan ahead ´

This is ridiculous.

 

[511]

Yeah let’s try to justify Intel’s ‘you will change your mobo every year’ by ´Intel is such incompetent that they didn’t thought maybe it could plan ahead ´

This is ridiculous.

You can look at pin layout lmao 1700 -> 1851 is not a small change also I said before they planned for 3 gens on LGA 1851 due to their skill issue we only got 1 gen they did plan ahead but messed it up.

 

[Khato]

Well, guess my impression wasn't incorrect and the 265k is pretty much 2x as fast at high quality AV1 transcode as the 14900k was. I'm not certain why my results show far more improvement than reviews. I know two primary differences are that I'm running notably higher quality settings and running on a full length 4k movie rather than a short clip. Maybe I'll torture the 7940HS HTPC with the same task at some point for another data point. Transcode time was 75354 seconds (2.89 fps) for the 14900k and 38364 seconds (5.68 fps) for the 265k.

For those curious, source was the 4k variant of the 3rd Harry Potter movie. Transcode done with Handbrake 1.9.2 at 3840x1600 output resolution using AV1 10-bit SVT encoder on preset 4 with multi-pass encoding at an average 14400 kbps along with an audio transcode to 16 bit FLAC.

 

[Doug S]

Zen 7 is not guaranteed to be AM5

What's the trigger for AMD to change sockets? Upgrading to DDR6 and/or PCIe6? Neither is happening in the mainstream market this decade (I would have said DDR6 will hit the mainstream at the very end of this decade, but the memory market chaos and AI bubble will push that back into the 2030s)

PCIe6 will be a lot more expensive because of how it works, and no one needs it for GPUs. If you need faster NVMe you can get PCIe NVMe RAID cards and run a bunch of them in parallel and go well beyond what a single PCIe6 NVMe drive would do. So I wouldn't hold my breath for PCIe6, because the use cases just aren't there in the AM5/AM6 market, and won't be for a LONG time..

 

[jdubs03]

Let’s be honest DDR5 and PCIe5 (same with WiFi 7) are plenty capable for the foreseeable future. In note of that, it’s possible that we could see adoption of the successor standards in around 2028. Though with DDR less likely with the ongoing disruption.

 

[511]

What's the trigger for AMD to change sockets? Upgrading to DDR6 and/or PCIe6? Neither is happening in the mainstream market this decade (I would have said DDR6 will hit the mainstream at the very end of this decade, but the memory market chaos and AI bubble will push that back into the 2030s)

DDR6 would be my best bet cause the memory bandwidth increase would be nice but I am expecting ddr6 to be in 28-29ish time

 

[Kryohi]

Well, guess my impression wasn't incorrect and the 265k is pretty much 2x as fast at high quality AV1 transcode as the 14900k was. I'm not certain why my results show far more improvement than reviews. I know two primary differences are that I'm running notably higher quality settings and running on a full length 4k movie rather than a short clip. Maybe I'll torture the 7940HS HTPC with the same task at some point for another data point. Transcode time was 75354 seconds (2.89 fps) for the 14900k and 38364 seconds (5.68 fps) for the 265k.

For those curious, source was the 4k variant of the 3rd Harry Potter movie. Transcode done with Handbrake 1.9.2 at 3840x1600 output resolution using AV1 10-bit SVT encoder on preset 4 with multi-pass encoding at an average 14400 kbps along with an audio transcode to 16 bit FLAC.

Did you power limit your 14900K to something like 65W? Those results are different from every review because you're doing something wrong.
And of course, a 9950x would be faster.

Intel Core Ultra 9 285K "Arrow Lake" Delivers Strong Linux Performance Review - Phoronix

www.phoronix.com

 

[DrMrLordX]

And of course, a 9950x would be faster.

It would also be more expensive than a 265k.

 

[Khato]

Did you power limit your 14900K to something like 65W? Those results are different from every review because you're doing something wrong.
And of course, a 9950x would be faster.

As previously stated, power limit was set to 200W.

Unlike reviews, my numbers are from actual use rather than single-task benchmarking. Default behavior with windows on the 14900k was to relegate handbrake to the e cores only if not left as the foreground window.

 

[LightningZ71]

DDR6 will be a necessity much sooner than PCIe 6.0. LPCAMM will allow faster data transmission and the DDR5 spec will run out long before the module will. Also, with now Intel, AMD and Apple pushing higher end iGPU configurations into the market, and NVidia maybe making enough market penetration some day to matter, motherboard installed DRAM performance will become more and more important. I also don't see the market turmoil stalling DRAM development. Instead, I see it pushing it faster. If AI proves not to be the bubble that many think it is, then it's demand for memory capacity and performance are insatiable.

As for PCIe 6.0, It's not enough bandwidth on x16 to make it enough for AI to stream data over, and there's no other application out there in consumer land with sufficient volume to warrant pushing it. Now, for server/HEDT/Workstation, that may not be entirely true, but it's going to add cost.

 

[eriksp92]

Does anyone here have an Arrow Lake mobile (H-series) chip and can run Cinebench 2026? My Zenbook 14 OLED with a 125H scores 2969, a bit more than the 2861 they have for their 255H reference system - with HT disabled, my system still scores 2705. Either that system has an absurdly low PL1 and PL2, or Intel 3 and 18A believers like me were just validated beyond belief, lol

 

[511]

Does anyone here have an Arrow Lake mobile (H-series) chip and can run Cinebench 2026? My Zenbook 14 OLED with a 125H scores 2969, a bit more than the 2861 they have for their 255H reference system - with HT disabled, my system still scores 2705. Either that system has an absurdly low PL1 and PL2, or Intel 3 and 18A believers like me were just validated beyond belief, lol

It all depends on the power limit it could be possible the ARL-H Score was done with very low PL

 

[poke01]

Does anyone here have an Arrow Lake mobile (H-series) chip and can run Cinebench 2026? My Zenbook 14 OLED with a 125H scores 2969, a bit more than the 2861 they have for their 255H reference system - with HT disabled, my system still scores 2705. Either that system has an absurdly low PL1 and PL2, or Intel 3 and 18A believers like me were just validated beyond belief, lol

There is one here

https://www.computerbase.de/forum/t...-gpus-in-cinebench-2026.2261907/post-31197126

 

[regen1]

CES 2026: Intel Core Ultra Series 3 Debuts as First Built on Intel 18A

www.intc.com

Some claims:

Core Ultra X9 and X7 Processors

Within the Intel Core Ultra Series 3 mobile lineup, a new class of Intel Core Ultra X9 and X7 processors come packed with the highest performing, integrated Intel® Arc™ graphics. They are purpose-built for multitaskers that handle advanced workloads like gaming, creation and productivity on the go. The top SKUs feature up to 16 CPU cores, 12 Xe -cores and 50 NPU TOPS, delivering up to 60% better multithread performance1, over 77% faster gaming performance2 and up to 27 hours of battery life3.

1 Up to 60% better multithread performance vs. Lunar Lake at similar power with an Intel® Core™ Ultra X9 388H vs. Intel® Core™ Ultra 9 288V. As measured by Cinebench 2024 Multi Core (at 25W). See intel.com/performanceindex for workloads and configurations. Results may vary.

2 Up to 77% better gaming performance vs. Lunar Lake with an Intel® Core™ Ultra X9 388H vs. Intel® Core™ Ultra 9 288V. As measured by Geomean of average game performance across 45 game titles at 1080p High with 2x upscaling when supported on Panther Lake Reference Platform vs. Lunar Lake Reference Platform. See intel.com/performanceindex for workloads and configurations. Results may vary.

3 Get as much as 27.1 hours while Netflix streaming with an Intel® Core™ Ultra X9 388H tested in Lenovo IdeaPad reference design. As measured by Netflix streaming at 1080p in Edge browser. Individual system results will vary significantly with different use, battery capacity and other factors. Learn more at intel.com/performanceindex.

 

[regen1]

Intel's Performance Index:

https://edc.intel.com/content/www/us/en/products/performance/benchmarks/intel-core-ultra-processors-series-3_1/

SKUs:





Edit: Adding another SKU list pic as there's some conflicting info