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.

 

[SiliconFly]

.

 

[OneEng2]

What's wrong with monts lol, ISA/avx512? That's solvable, especially if they redesign it as a full fledged P core. Lack of SMT? New Coves don't have it either. Arrow is not the only implementation of the P core. Lunar lake has the fixed version without latency problems, yet still lags Apple, QC Oryon 2, arm x925 on a similar node.

I believe that new Coves for server have SMT while new Monts (even for server) don't.

If you add everything that Cove enjoys to mont, is it still an E core? Is it still "mont"?

Good point on Lunar Lake though. It essentially enjoys all the benefits of a monolithic die.

Do you have a few benchmarks to make this point?

 

[igor_kavinski]

Do you have a few benchmarks to make this point?

I do!

ASUSTeK COMPUTER INC. ASUS Zenbook S 14 UX5406SA_UX5406SA_010318305 vs Dynabook Inc. PORTEGE X40L-M - Geekbench

288V holds up really well despite having only 8 cores.

 

[coercitiv]

I personally think (here is a counter prediction) that it is more likely that Cove will become the backbone single unifying architecture and that "mont" will go away .

A number of people around here tend to focus more on the core itself, as if it's just a solution to a physics problem. A few other people, including some with insider insights, bring up the human factor: the cores themselves are also the manifestation of a team's mentality and influence inside the company, with some decisions dictated by politics rather than engineering logic.

No matter what you or I believe about the Coves, there's undeniable evidence out there that Intel's big cores have been slacking hard in the last decade. Arguing that the Cove will become the backbone of a unified arch is either a purely subjective take, or one based on the belief that the Cove team will "kill" or sideline the Mont team.

Based solely on engineering merit, the Mont team has earned their chance to attempt a flagship core. Whether that materializes at Intel... I do not know.

If you add everything that Cove enjoys to mont, is it still an E core? Is it still "mont"?

It's still a Mont. It may no longer be an E core though. Give it it's own cache to trash, moar AVX, optimize +10% fmax and PPA is gone now, it's a semi-P core. I'm sure the Mont team has better ideas than that though.

 

[igor_kavinski]

Based solely on engineering merit, the Mont team has earned their chance to attempt a flagship core. Whether that materializes at Intel... I do not know.

Sadly, we only find out about their success or failure in 2028, or worse, 2029!

 

[SiliconFly]

.

 

[cannedlake240]

Sadly, we only find out about their success or failure in 2028, or worse, 2029!

No need for such long wait probably. By then Intel might be reduced to being a trailing edge foundry making legacy chips for a few clients or a computer history museum lol

 

[OneEng2]

I do!

ASUSTeK COMPUTER INC. ASUS Zenbook S 14 UX5406SA_UX5406SA_010318305 vs Dynabook Inc. PORTEGE X40L-M - Geekbench

288V holds up really well despite having only 8 cores.

Unfortunately, these benchmarks do not isolate either core type to make the point clear (or not). Additionally, I think real applications might be better than a benchmark application, but even the benchmark apps would need to have the cores isolated to know much about the relative usefulness of P vs E cores.

A number of people around here tend to focus more on the core itself, as if it's just a solution to a physics problem. A few other people, including some with insider insights, bring up the human factor: the cores themselves are also the manifestation of a team's mentality and influence inside the company, with some decisions dictated by politics rather than engineering logic.

No matter what you or I believe about the Coves, there's undeniable evidence out there that Intel's big cores have been slacking hard in the last decade. Arguing that the Cove will become the backbone of a unified arch is either a purely subjective take, or one based on the belief that the Cove team will "kill" or sideline the Mont team.

Based solely on engineering merit, the Mont team has earned their chance to attempt a flagship core. Whether that materializes at Intel... I do not know.


It's still a Mont. It may no longer be an E core though. Give it it's own cache to trash, moar AVX, optimize +10% fmax and PPA is gone now, it's a semi-P core. I'm sure the Mont team has better ideas than that though.

I am not disagreeing with you .... or not completely. I am pointing out that Cove has to handle challenging loads that mont is being shielded from. Cove has to deal with Intel's move to tiles and the associated latency problems that introduces. Cove has also undergone a major decrease in power consumption vs RPL. We all do still remember that RPL parts literally cooked themselves ..... right? Intel could hardly continue on that path. I think we can all agree the "power be damned" train at Intel ran totally and completely out of tracks.

I don't disagree that mont has done some pretty cool things (pun intended). Still, I maintain that mont will perform miserably as-is on many workloads that Cove is currently doing quite well on.

Still, what we CAN compare fairly easily is the PPA of a Zen 5 full core to the PPA of a full Cove core. In that comparison, it is clear that Cove is having difficulties besting a core that is on a lesser node and is quite large compared to a Zen 5 on a similar node.... so yea, I think they have some work to do.

And mont team did some pretty amazing things as well, I just find it amusing that the same people that curse ARL for its poor gaming performance, think getting rid of Cove is a good idea as I believe that mont would be a terrible gaming CPU.

 

[DavidC1]

And mont team did some pretty amazing things as well, I just find it amusing that the same people that curse ARL for its poor gaming performance, think getting rid of Cove is a good idea as I believe that mont would be a terrible gaming CPU.

It can't be that terrible, as 1P+16E performed better than the default 8P+16E config in many gaming workloads. At some point you have to make a conclusion that their P core/team simply sucks.

It may not make sense, but think of current hybrid as if the "P" cores were based on Prescott 3.4GHz and "E" cores were based on Yonah 2GHz.

Pentium 4 also did pretty well against Pentium 3 in many workloads.

I would prefer they make it 16+8. Eight weak cores are enough for background stuff.

16+2 looking at the way P and E core development is going and we'd be fine with it.

 

[itsmydamnation]

We were talking about when he was ceo. Like Lisa.

Its not equivalent, Lisa spent her entire career in micro electrics , Pat spent over a decade at vmware and that decade has seen a lot of change in the way other ( not intel) operated and approached designing micro processors. You cannot assume Pat had the same insight or understanding Lisa has and visa versa.

But more importantly AMD's turn around was more culture not individuals ( individuals can certainly destroy a culture but they rarely make one) , you can see from the way Jim Kellar talks about the start of Zen , their engineers had a loosing culture. Is intels culture being fixed , can they continue to off set bad culture by lighting money on fire etc.

 

[511]

Its not equivalent, Lisa spent her entire career in micro electrics , Pat spent over a decade at vmware and that decade has seen a lot of change in the way other ( not intel) operated and approached designing micro processors. You cannot assume Pat had the same insight or understanding Lisa has and visa versa.

We can't really compare them Pat has more technical Feats while Lisa has better Feats as a CEO granted she didn't have to manage both Fabs and Design

But more importantly AMD's turn around was more culture not individuals ( individuals can certainly destroy a culture but they rarely make one) , you can see from the way Jim Kellar talks about the start of Zen , their engineers had a loosing culture. Is intels culture being fixed , can they continue to off set bad culture by lighting money on fire etc.

This is something only time can answer

 

[itsmydamnation]

We can't really compare them Pat has more technical Feats while Lisa has better Feats as a CEO granted she didn't have to manage both Fabs and Design

This is something only time can answer

You know Lisa was an actual semi conductor researcher. All things begin equal she probably have a better handle on managing Fabs then Pat.

also TIL that Lisa ran the Team at IBM who came up will Cell.

 

[cannedlake240]

believe that new Coves for server have SMT while new Monts (even for server) don't.

Most rumors suggest otherwise

If you add everything that Cove enjoys to mont, is it still an E core? Is it still "mont"?

That should be the goal of a new core project, redesign atom for higher performance so it can replace both cove and monts. Intel supposedly doesn't want to finance multiple CPU designs when it's AI and foundry that are becoming more relevant.

Do you have a few benchmarks to make this point?

In any M4, Oryon 2 review it's evident Lunar is behind in efficiency and even peak 1T performance. This shouldn't be the case if Intel is to continue competing

 

[511]

You know Lisa was an actual semi conductor researcher. All things begin equal she probably have a better handle on managing Fabs then Pat.

also TIL that Lisa ran the Team at IBM who came up will Cell.

You seems to forget Pat was the chief architect of i486 and lead part of key technologies like USB/Wi Fi/Core/Xeon.In terms of feat in tech he out classes her as for handling fabs the reason for AMDs success is in part due to TSMC as well they have outsourced one of the worst business to run in terms of 💰 also running fabs now is more suicidal than it was 10-15 years ago

 

[coercitiv]

the reason for AMDs success is in part due to TSMC

Reading this felt like a Picard head in hands moment. Intel is using TSMC as we speak. Where's the success?! Stop looking outside for reasons this company is failing. The reasons are inside. Intel inside.

 

[Nothingness]

You seems to forget Pat was the chief architect of i486 and lead part of key technologies like USB/Wi Fi/Core/Xeon.In terms of feat in tech he out classes her as for handling fabs the reason for AMDs success is in part due to TSMC as well they have outsourced one of the worst business to run in terms of 💰 also running fabs now is more suicidal than it was 10-15 years ago

Addressing just the CPU part of your comment, having designed a CPU released in 1989 gives exactly no clue how current CPUs are being designed and what features are needed to make them fast/efficient/validated/... His old knowledge is almost completely irrelevant.

 

[SiliconFly]

.

 

[DavidC1]

Addressing just the CPU part of your comment, having designed a CPU released in 1989 gives exactly no clue how current CPUs are being designed and what features are needed to make them fast/efficient/validated/... His old knowledge is almost completely irrelevant.

Reading this felt like a Picard head in hands moment. Intel is using TSMC as we speak. Where's the success?! Stop looking outside for reasons this company is failing. The reasons are inside. Intel inside.

I think the problem with Intel is long-standing and would take almost as long to fix, because the cultural problems existed all the way back in the 90's. It pretty much existed ever since the original three left Intel.

The last time when Pat could have improved it was if he was Otellini's successor, back when Pat was CTO. Kraznich screwed up the company badly enough that I wonder if it's even doable at this point. You probably need firing 60-70% of employees - basically start over to kill all the internal sabotaging between teams.

 

[511]

Reading this felt like a Picard head in hands moment. Intel is using TSMC as we speak. Where's the success?! Stop looking outside for reasons this company is failing. The reasons are inside. Intel inside.

Intel's troubles are all inside I wouldn't argue that as for success we have to see the promised 18A cause

I didn't get the Picard Head in hand moment reference

I think the problem with Intel is long-standing and would take almost as long to fix, because the cultural problems existed all the way back in the 90's. It pretty much existed ever since the original three left Intel.

The last time when Pat could have improved it was if he was Otellini's successor, back when Pat was CTO. Kraznich screwed up the company badly enough that I wonder if it's even doable at this point. You probably need firing 60-70% of employees - basically start over to kill all the internal sabotaging between teams.

60% is tooooo much more like 30% on top of what they laid off they have manufacturing as well so we need to keep that in mind

 

[DavidC1]

60% is tooooo much more like 30% on top of what they laid off

Intel's problems were serious back in the 90's. Basically the feeling I got was they retained most of Grove's negative traits while losing the part what made it work despite his flaws.

I say 60-70% because it's what would take to "break" that bad culture. A sweeping change. Maybe even more. The aftermath would result in something that some would no longer classify as Intel. The difficulty is making the technicals work after that sweeping change of course, because the culture is ultimately responsible for a company's rise to the top, and you'd need a completely new recipe.

A bit of a history lesson. Successor to Otellini was originally not Kraznich but likely Sean Maloney, but in 2012 he had to resign due to a stroke:

Intel Executive Vice President Sean Maloney to Retire

www.intc.com

Sean Maloney: The Man Who Could Have Led Intel

Sean Maloney (above) spent nearly three decades being groomed for Intel's top job then he suffered a devastating stroke. Here's his remarkable story of recovery.By Intel Free PressOn the car ride home from the hospital after suffering a stroke in February 2010, Intel executive Sean Maloney...

www.siliconvalleywatcher.com

He doesn't have a comprehensive engineering background, but it still is in computer science, trained under Grove, and knew to a degree that mobility was the future.

Would have been WAY better than Kraznich. The reason they lost Mooly Eden, Dadi Perlmutter, all who led Pentium M effort was because Kraznich kicked out anyone who stood against him.

It's no point talking about it. Companies are made of living people and thus reflect their goals, their aspirations, and their flaws. It eventually dies too.

 

[igor_kavinski]

Totally. Lisa knows all about microprocessors and he doesn't.

She's actually more of a people person which helps when you are running a huge company filled with, you know, people

Had Pat been a good Christian (not arrogant), he would have done well. He was too arrogant in his statements and he was too arrogant to think that God was on his side and that he could do no wrong. He forgot the simplest of things. God's Universe is running on Auto based on His laws and there is simply NO substitute for hard work, humility, thinking HARD, being true to yourself, actually caring about the people working for you and demanding the absolute best from your teams because no one is going to accept any excuses when those teams fail you and it's your head on the platter.

 

[DavidC1]

She's actually more of a people person which helps when you are running a huge company filled with, you know, people

AMD was much smaller(still is), the bureaucratic part not infesting the company, nor did they have decades of toxic culture.

Turning around was much, much easier than Intel.

 

[igor_kavinski]

AMD was much smaller(still is)

Well, guess what? Intel's toxic culture was maybe the best thing ever to happen to them because looks like they TOO will soon be a small company

 

[511]

The only guy capable to do turn around of that scale is dead unfortunately

 

[igor_kavinski]

also running fabs now is more suicidal than it was 10-15 years ago

Yes, all the fabs other than TSMC that are doing just fine and dandy are committing suicide every single day