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.

 

[Magio]

All things considered it seems like ARL and Zen 5 are gonna be pretty evenly matched in most ways. But Intel is coming in later and using a more expensive node than they won't have infinite capacity on. I don't think either side is gonna eclipse the other on this cycle, which is a disappointment both ways: This was meant to be Intel's big return to form with an (almost) leading edge node and Zen5 was long touted as the biggest IPC gain in x86 in ages. Ultimately it will be a modest improvement for both.

Intel does have LNL which is set to stand out in its own category, but how much it will help them will largely depend on the design wins it gets.

 

[SiliconFly]

-

 

[mikk]

Lunar Lake doesn't have a hardware encoder or is it because it's not supported yet? Only decoding is listed here.

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

 

[poke01]

All things considered it seems like ARL and Zen 5 are gonna be pretty evenly matched in most ways. But Intel is coming in later and using a more expensive node than they won't have infinite capacity on. I don't think either side is gonna eclipse the other on this cycle, which is a disappointment both ways: This was meant to be Intel's big return to form with an (almost) leading edge node and Zen5 was long touted as the biggest IPC gain in x86 in ages. Ultimately it will be a modest improvement for both.

Intel does have LNL which is set to stand out in its own category, but how much it will help them will largely depend on the design wins it gets.

On desktop it will ultimately come to gaming performance, platform longevity and price.

AMD may have the lead in all three when the x3D chips launch but for the niche like creatives they may use Intel.

 

[poke01]

At the same time, the power consumption of one core has increased more than two times despite being on a much newer node. That's the price Apple pays to have a wider design. The power consumption scales much faster than with simpler Intel and AMD designs

This isn’t true, Apple still uses much less power in ST despite having 10-wide decoder while delivering much higher IPC than AMD/Intel. The reason why power consumption and clock increased is because of the HP cell for the P cores. The 10watt is for 1 thread and using than more 3 threads drops to 3.94GHz which should halve the power consumption.

Clocks do matter but within reason. As you will see with lunar, Intel won’t achieve same ST as Apple and power consumption for ST will be around 15 watts at 5.1GHz despite having a less wider design( this is what I think. It will have to tested). The all core load clock will be much lower and this is why Lunar scores low in MT compared to M4 in the scores we got so far.

If we compare with AMD, AMD Strix Point (2x4 decoder) around 20 watts at 5.1GHz in CB 2024 ST and scores lower than M3(9-wide decoder) at 4.05GHz which uses 10 watts. We will have to wait and see how M4 compares.

 

[jdubs03]

Just going by the spec2017 results, it’s possible that the M4 could use 15W to achieve its max score. It does use a good bit more power than the M3 p-core.

 

[poke01]

Just going by the spec2017 results, it’s possible that the M4 could use 15W to achieve its max score. It does use a good bit more power than the M3 p-core.

M4-P core is definitely tuned for performance instead of efficiency like M3.
Doesn’t Intel also use HP cells?

Edit: looks like Intel also uses HP cells for its nodes and cores.

 

[jdubs03]

Pretty well known for it I think.

Btw I was just going by Geekerwan results show the performance and power draw results of the M4 vs M3 and applied it to the 10W you noted for the M3 in CB 2024 ST. I’m assuming notebookcheck was the source for that?

 

[poke01]

I’m assuming notebookcheck was the source for that?

Computer base.

 

[DavidC1]

Clocks do matter but within reason. As you will see with lunar, Intel won’t achieve same ST as Apple and power consumption for ST will be around 15 watts at 5.1GHz despite having a less wider design( this is what I think. It will have to tested).

If Lunarlake achieves great battery life in general, that would mean they reduced their "not core" power. In that case, peak power use in all cases will be reduced by an identical amount.

It won't be much, but 2W reduction would result in 13W.

I don't think either side is gonna eclipse the other on this cycle, which is a disappointment both ways: This was meant to be Intel's big return to form with an (almost) leading edge node and Zen5 was long touted as the biggest IPC gain in x86 in ages. Ultimately it will be a modest improvement for both.

What matters is the end result. If they perform similar and Intel gets power use down then it really doesn't matter.

Strix Point tests show the "same core" approach still has the same scheduling problems and in some ways are behind Intel because it's their first implementation.

Leading edge doesn't really help super-high clocked chips approaching 300W. It helps in die area and in low power parts like Lunarlake.

 

[poke01]

On If Lunarlake achieves great battery life in general, that would mean they reduced their "not core" power. In that case, peak power use in all cases will be reduced by an identical amount.

It won't be much, but 2W reduction would result in 13W.

Yes Lunar Lake will definitely achieve great battery life and Intel did many things to the Lunar lake design do so. It’s the most advanced x86 SoC mobile Intel has ever created.

 

[SiliconFly]

-

 

[DavidC1]

Yes Lunar Lake will definitely achieve great battery life and Intel did many things to the Lunar lake design do so. It’s the most advanced x86 mobile Intel ever created.

The thing is they had the formula more than 10 years ago with the Atom parts, but gave up on the system-level optimization needed for the amazing battery life.* It happened on Braswell generation. They should have kept it.

*6-8 hour battery life on a 8-inch screen with a 15WHr battery. Idled screen on at 1.3-1.5W.

 

[poke01]

Yes Lunar Lake will definitely achieve great battery life and Intel did many things to the Lunar lake design do so. It’s the most advanced x86 SoC mobile Intel has ever created.

To add to this Intel still suffers from Lion Cove, I bet if Intel had created an x86 p core with very high IPC and lower clock it would have been on par with the M3 in all areas.

They got rid of SMT in lunar and when will Intel get rid of 8/16/32bit, who knows?*

* I mean when will they release such a design

 

[jdubs03]

Computer base.

I checked two different articles and they both have a consistent CPU package power of around 4.5-6W:

Apple MacBook Air mit M3 in 15 Zoll im Test

Nach mehreren Tagen mit dem neuen, M3-basierten MacBook Air 15" lässt sich eine eindeutige Kaufempfehlung für dieses Notebook aussprechen.

www.computerbase.de

^shows a quick spike in the beginning of the test to 10.6W, then decreases from 5.5W to 4.5Wish toward the end of the benchmark.

Apple MacBook Pro 14" mit M3 Max im Test

Apple macht das bereits sehr gute MacBook Pro durch einen überarbeiteten Bildschirm und die neuen M3-Prozessoren noch ein Stückchen besser.

www.computerbase.de

^max is 6.2W, but slowly decreases to 5.1W at the end.

NBC has single-core load at 6.5W dropping to 5.5W:

Apple M3 SoC analyzed: Increased performance and improved efficiency

A Notebookcheck analysis of the performance and efficiency of the new Apple M3 SoC, Apple's first 3nm chip, in comparison with AMD, Intel and Qualcomm.

www.notebookcheck.net

So really that reflects even better on Apple. And lowers my estimate for the M4 to consistent around 8W single-core load. Big big difference to saying 15W before. The gap is huge.

 

[DavidC1]

BTW there's a reason why Intel was having a hard time with their first try as a foundry: their design flow were a match for their internal CPU teams; anything designed differently, notably with process portability in mind, didn't work. I don't know where they stand now, but given that they use TSMC, I guess they have raised the bar. But Arm and other IP providers have decades of experience here.

That's an understatement. Now we know when Otellini was talking about foundry efforts, it was merely a gesture and nothing more. They thought of potential customers as an addition, a cherry on top of their revenue and nothing else.

If what Semiwiki said was right, then Intel even treated equipment vendors badly. That's what happens when pride and hubris is unchecked after unmatched dominance.

I thought their mobile loss in 2011 was bad. The situation now is much worse compared to back then. Now their foundation is crumbling. People talk about needing to separate the company into two, but right now there's question on the capabilities of both sides. It's a fab still behind years coupled with design team that's behind even average.

So really that reflects even better on Apple. And lowers my estimate for the M4 to consistent around 8W single-core load. Big big difference to saying 15W before. The gap is huge.

The ARM leadership is so great it even makes me question whether Intel had a top development team with Conroe. I'm thinking they merely had the attention, but not actual technical capability. The advent of Smartphones really, really accelerated development.

 

[Abwx]

I checked two different articles and they both have a consistent CPU package power of around 4.5-6W:

Apple MacBook Air mit M3 in 15 Zoll im Test

Nach mehreren Tagen mit dem neuen, M3-basierten MacBook Air 15" lässt sich eine eindeutige Kaufempfehlung für dieses Notebook aussprechen.

www.computerbase.de

^shows a quick spike in the beginning of the test to 10.6W, then decreases from 5.5W to 4.5Wish toward the end of the benchmark.

Apple MacBook Pro 14" mit M3 Max im Test

Apple macht das bereits sehr gute MacBook Pro durch einen überarbeiteten Bildschirm und die neuen M3-Prozessoren noch ein Stückchen besser.

www.computerbase.de

^max is 6.2W, but slowly decreases to 5.1W at the end.

NBC has single-core load at 6.5W dropping to 5.5W:

Apple M3 SoC analyzed: Increased performance and improved efficiency

A Notebookcheck analysis of the performance and efficiency of the new Apple M3 SoC, Apple's first 3nm chip, in comparison with AMD, Intel and Qualcomm.

www.notebookcheck.net

So really that reflects even better on Apple. And lowers my estimate for the M4 to consistent around 8W single-core load. Big big difference to saying 15W before. The gap is huge.

M3 Max does 23800 pts in CB R23 at 58-60W while AMD s 370 does 23000 pts at 65W, overall efficency is comparable, that s more relevant than ST scores, so x86 is as efficient as Apple CPUs at comparable perf.

Beside M3 Max has 16 cores while the 370 has only 12, at 16C the latter would be more efficient, moreover if shrinked to the same N3 process as the former.

 

[jdubs03]

M3 Max does 23800 pts in CB R23 at 58-60W while AMD s 370 does 23000 pts at 65W, overall efficency is comparable, that s more relevant than ST scores, so x86 is as efficient as Apple CPUs at comparable perf.

Beside M3 Max has 16 cores while the 370 has only 12, at 16C the latter would be more efficient, moreover if shrinked to the same N3 process as the former.

At its highest point it reached 59W, but then it’s decreasing as time goes on. The average is thus much lower, around 45W. I suppose if the power draw could be sustained, the score would be a decent bit higher.

Cinebench R23 Multi

TDP	Ryzen AI 9 HX 370	Ryzen Z1 Extreme
15 Watt	10,435 points	8,635 points
20 Watt	12,627 points	10,798 points
28 Watt	15,849 points	13,002 points
	Ryzen AI 9 XH 370	Ryzen 9 8945HS
35 Watt	17,990 points	14,423 points
45 Watt	20,113 points	15,506 points
55 Watt	21,625 points	16,482 points
65 Watt	22,960 points	17,077 points

At 60W, HX 370 is around 22,293.
Since the benchmark is a sustained workload we shouldn’t be using the max value. But rather the average. At 45W the HX 370 does around 20110-20480 (Geekerwan).
23807 vs 20480 at the same power level is 16pct.

But you’re also not including that the HX 370 released over 9 months later. It’s main competition through its lifetime will be the M4 Max.
And Zen5-R against M5-series.

I will note though, based on GB5 multicore score, the M4 doesn’t really see an improvement in terms of pts/core compared to M3; comparing the iPad Pro to a MacBook Air (closest comparison I could find with fanless devices, IPadOS vs MacOS is barely different in results (1pct higher for MacOS)).
And from what I see in those results the pts/core even within the M3 series goes down as the core count increases. Need CB scores to get a better handle of that behavior.
But obviously based on what we have right now, the M4 is not an improvement over M3 in terms of multicore performance. Will be good to see if that remains the case once the MacBooks come out. But based on what I’ve just noted here, at least in GB5 the M4 Pro/Max would be the same as the M3 Pro/Max EDIT: (as long as core counts remain the same).

 

[DavidC1]

But you’re also not including that the HX 370 released over 9 months later. It’s main competition through its lifetime will be the M4 Max.
And Zen5-R against M5-series.

Great thing about NBC is you can find the CPU you want it to compare exactly with it. Type in M3 Max and then Ryzen AI 9 HX 370.

NBC comparison shows M3 Max has 35% higher efficiency compared to Ryzen AI 9 HX 370, and it's at the system level sans LCD, so the actual advantage will be higher.

35-45% advantage is at least two full generation of gains, and it's in multi-core where it's easier to catch up, much much harder than in per thread performance. And based on cadence of the two x86 manufacturers, two generation = 3.5-4 years.

Let's put it this way, if the tables are reversed and it's AMD having that lead with Apple having high clocks, then AMD folks would be shouting from the rooftops.

 

[desrever]

Let's put it this way, if the tables are reversed and it's AMD having that lead with Apple having high clocks, then AMD folks would be shouting from the rooftops.

I like how you bring this up in a thread about Intel products. You literally can't stop talking about Apple vs AMD now, how close is it from shouting from rooftops do you want to get on a tech forum?

 

[DavidC1]

I like how you bring this up in a thread about Intel products. You literally can't stop talking about Apple vs AMD now, how close is it from shouting from rooftops do you want to get on a tech forum?

Right, cause I'm the only one talking about it.

 

[poke01]

I like how you bring this up in a thread about Intel products. You literally can't stop talking about Apple vs AMD now, how close is it from shouting from rooftops do you want to get on a tech forum?

Agree, no need split into arguments into fanbases, gets ugly. Best stick to technical comparisons only.

 

[poke01]

M3 Max does 23800 pts in CB R23 at 58-60W while AMD s 370 does 23000 pts at 65W, overall efficency is comparable, that s more relevant than ST scores, so x86 is as efficient as Apple CPUs at comparable perf.

Beside M3 Max has 16 cores while the 370 has only 12, at 16C the latter would be more efficient, moreover if shrinked to the same N3 process as the former.

Its looks as though ARM cores don't as well in R23 compared to CB 2024. There are considerable improvements in CB 2024.

Cinebench 2024: Reviewing the Benchmark

Maxon’s Cinebench is a perennial benchmark favorite.

chipsandcheese.com

If we take a look at CB 2024 the %'s changes. The M3 Max takes a bigger lead, yes it has 12P+4E core layout but it also has no SMT. So 16 threads vs 24 threads.

The more fair comparison would be the M3 Pro vs Strix point both have 12 cores but m3 pro has 12 threads and Strix has 24 threads. At simliar wattage they are score simliarly.



It would interesting to see how Arrow Lake H fares with the loss of HT. Intel massively improved the e-cores to the point where MT increased despite the HT loss.

 

[LightningZ71]

I love how they forget that M3 is on TSMC N3b and M4 is on N3e with finflex as compared to Strix Point being on N4p when they do these power/performance comparisons. N3 family is roughly 25-30% better at power/iso perf than N5 family, which N4 is a mild improvement on. That Strix falls in that window is a testament that they are, from a core design point of view, roughly equivalent at perf/watt with notably different approaches and different ISAs. Supposedly, Intel 3 is in the same ballpark as N4p as well, maybe leading on feasible density.

The real story is that Apple is aiming at a more premium market with the M series, so they can afford the higher per chip cost of leading edge nodes with respect to AMD and Intel on the desktop and mobile.

When we get final Lunar Lake nunbers vs. M3, we'll have a better temperature of this market. M4 should still maintain a decent lead due to being a node tweak or more ahead of the rest.

 

[jdubs03]

I’ve been mulling this over a bit. Correct me if off base.

But wouldn’t the best way to measure multicore performance is for example take the CB23/24 nT score in points then divide by clock speed (all-core) then by thread count and then by wattage (or exclude wattage if normalized)? I assume area would be a nice variable in there too but I feel like that’s tougher to get data on. (This would also apply to single core, but we aren’t talking about that right now, this is definitely where Apple is way ahead).

I was thinking about M3 -> M4 and Strix Point, and by using the metric of pts/clock/thread, both actually went down compared to their predecessors.

If the lunar lake is able to maintain similar multicore performance with those less threads that would be a real solid improvement, particularly at iso-power (or even better less power draw).