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.

 

[poke01]

Also, there are some results for Geekbench 5. The P-core in Intel Core 285K matches the one in Apple M4:

That M4 score is throttled. Here's the non-throttled one. Most of the M4 scores will be lower because you know thin tablet, it peaks at around ~2700-2750 in GB5.
There is a difference when the frequency is maintained throughout the test.
https://browser.geekbench.com/v5/cpu/compare/22615013?baseline=22798191

 

[Abwx]

Nope. Check the videocardzGB6 numbers (all stock). 9950X is 3335 & 285K is 3450. That puts 285K around 3.5% ahead of 9950X.

I posted Computerbase since all their tests are at stock 5600 RAM using the latest AGESA contrary to random GB subsmissions whose most were done before the CPU was launched and with old AGESAs, hence the low average.

Beside even for the 14900K at 5600 RAM they have a higher number than the average GB submissions that often use higher speed RAM but are done by PCs that run some light background tasks when tested by random users.

 

[branch_suggestion]

Status quo, LNL is the only Intel part that actually makes a difference in their footing.

 

[TwistedAndy]

That M4 score is throttled. Here's the non-throttled one. Most of the M4 scores will be lower because you know thin tablet, it peaks at around ~2700-2750 in GB5.
There is a difference when the frequency is maintained throughout the test.
https://browser.geekbench.com/v5/cpu/compare/22615013?baseline=22798191

Here are a dozen of the latest Geekbench results for Apple M4. I don't see 2700-2750 there. The regular scores are closer to 2500-2600.

2700-2750 were probably obtained under LN as Geekerwan did.

 

[Henry swagger]

Very interesting! But I don't think it's fair to compare a sub-28W M4 with a 285K that runs at excess of 100W or something.

Just comparing ipc.. apple is the main competition with intel in terns of ipc

 

[Henry swagger]

Here are a dozen of the latest Geekbench results for Apple M4. I don't see 2700-2750 there. The regular scores are closer to 2500-2600.

2700-2750 were probably obtained under LN as Geekerwan did.

Matching apple p core with just 500 robs.. cougar cove will have even higher ipc 🙂

 

[poke01]

2700-2750 were probably obtained under LN as Geekerwan did.

Most likely the freezer. Almost all of results in the screenshot the freqency is not maintained and if it is maintained it would score around 2700. This shouldn't be an issue for Macs with fans.
https://browser.geekbench.com/v5/cpu/22615013.gb5

Spoiler: Freq for M4 - non-throttled

},
"processor_frequency": {
"frequencies": [
4407,
4405,
4394,
4399,
4392,
4402,
4402,
4402,
4408,
4396,
4401,
4406,
4399,
4396,
4406,
4402,
4338,
4403,
4398,
4405,
4397,
4396,
4406,
4402,
4397,
4406,
4403,
4407,
4402,
4397,
4401,
4401,
4402,
4408,
4402,
4390,
4398,
4403,
4407,
4407,
4403,
4404,
4391,
4371
]
},

The ones that are throttled score around ~2500.

Geekbench Browser

Spoiler: M4 - throttled

"processor_frequency": {
"frequencies": [
4388,
4401,
4394,
4392,
4380,
4393,
4401,
4409,
4395,
4390,
4394,
4398,
4391,
4403,
4393,
4398,
4381,
4384,
4394,
4383,
4388,
4388,
4386,
4330,
4386,
4396,
4395,
4396,
4400,
4396,
4394,
4382,
4393,
4389,
4396,
4381,
4389,
4399,
4385,
4395,
4394,
4395,
4395,
4399
]
},

 

[poke01]

Matching apple p core with just 500 robs.. cougar cove will have even higher ipc 🙂

Yeah I wonder how Intel's P cores will evolve, Intel did say they were focused on IPC instead of clock improvements now. But keep in mind that is 4.40GHz vs 5.7GHz Intel has a long way too go to match Apple in IPC.

At this rate the E-core team will surpass the P core.

 

[SiliconFly]

.

 

[Henry swagger]

Yeah I wonder how Intel's P cores will evolve, Intel did say they were focused on IPC instead of clock improvements now. But keep in mind that is 4.40GHz vs 5.7GHz Intel has a long way too go to match Apple in IPC.

At this rate the E-core team will surpass the P core.

Yeah per clock ipc apple is way ahead.. i think darkmont or arctic wolf will catch apple on a per clock ipc

 

[Abwx]

Thats invalid. Test 285K at recommended settings including RAM (not 5600).

Hence, 285K is 3.5% ahead of 9950X. No doubts.

Latest AGESA was received by Computerbase 4 days before the reviews days, actually they only had 3 days to do the tests, so all others AGESA, thats is, the totality of what was submited at GB before August 14, are invalid by the definition.

That being said i m curious of your reaction once some other submissions of the 285K are done at GB and that some scores are lower, because a single test is not representative statistically wise.

 

[511]

Yeah per clock ipc apple is way ahead.. i think darkmont or arctic wolf will catch apple on a per clock ipc

Darkmont won't it is similar to Crestmont vs Gracemont so i think 5-7% IPC on a even better node considering TSMC Said 18A rivals N3P

 

[TwistedAndy]

Most likely the freezer. Almost all of results in the screenshot the freqency is not maintained and if it is maintained it would score around 2700. This shouldn't be an issue for Macs with fans.
https://browser.geekbench.com/v5/cpu/22615013.gb5

There is something wrong with those clocks because the averages are 4400 for the first result and 4390 for the other ones. It's a 0.2% difference.

So, most results between 2500 and 2600 are valid and not throttled.

There are some weird scores (less than 2400 and more than 2700), but they are not representative and should be excluded from the analysis as outliers (if we want to stay objective, of course).

 

[SiliconFly]

.

 

[SiliconFly]

.

 

[Abwx]

So be it.

You answered before i ended editing my post....

 

[SiliconFly]

.

 

[Henry swagger]

Ok. It's time for me prepare a "I told you so" list of all the people who said this ARL generation will only get a lower single digit % performance uplift in ST. Whereas, it's already showing 11.7% ST gains against previous gen equivalent 14900K and this isn't final silicon yet!

There are many people in the list, with whom shall I start?

Mild said arrow lake refresh will uave even higher clocks

 

[KompuKare]

Yeah per clock ipc apple is way ahead.. i think darkmont or arctic wolf will catch apple on a per clock ipc

Surely the "per clock" is totally irrelevant when talking about IPC (Instructions Per Cycle).

While the x86 vendors get close to Apple they do so use way more power.

 

[cebri1]

Mild said arrow lake refresh will uave even higher clocks

Now we cannot rule out clock regression

 

[Magio]

Surely the "per clock" is totally irrelevant when talking about IPC (Instructions Per Cycle).

While the x86 vendors get close to Apple they do so use way more power.

For single core that continues to be true however LNL will be pretty close to a M3 in MT scenarios at comparable power, on the same node. (M4 will be ahead but on N3E instead of N3B.)

 

[poke01]

this is just Geekbench, we still have to test many more. Cinebench. Productivity tests, web tests, gaming etc

Edit: skymont is star here

 

[SiliconFly]

.

 

[Hulk]

this is just Geekbench, we still have to test many more. Cinebench. Productivity tests, web tests, gaming etc

Edit: skymont is star here

Looks like it's everything Intel said it was. Can't wait for actual tests with real software. Top of the AMD/Intel stack will once again be neck-and-neck in performance.

 

[SiliconFly]

.