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.

 

[jpiniero]

gaming excluded

But gaming is what drives DIY.

 

[511]

let gamer's buy a worse general purpose CPU than

 

[Hitman928]

let gamer's buy a worse general purpose CPU than

They often do, which is why not even AMD’s best general purpose CPU is the best seller and why i3s used to dominate any AMD CPU, even though their productivity benchmarks were weaker in general.

On the business side it’s more about existing relationships, service contracts, total costs for the “fleet” of computers, and what the sales person can convince the company they need (or that they’re getting a good deal).

 

[Hulk]

It's a good product now, after numerous firmware updates and a considerable price drop.

When it launched, the 265K was just 2% faster than 14700K in Techpowerup's testing (performance in applications, gaming excluded). The worst offender in my book was the poor browser performance, making it slower than a 12700K in Speedometer or Jetstream. The pricing added insult to injury, as it matched the 9900X.

View attachment 132698

People bash Arrow Lake because it started on the wrong foot and Intel never did manage to shake this bad image.

The 285K was selling for $420 at Microcenter a few weeks ago. That nuts for the top of the line part. I was tempted to buy one just to fool around with it.

 

[Khato]

If you design a processor to run at 6GHz on a certain process specification, but the process ends up at 85% of spec, then the processor will only run at 5.1GHz. So if PTL was designed on a PDK that promised 1.4x improvement over N3B, but the process only delivered a 1.2x improvement over N3B you'd still have a better process than N3B despite the lower clock speed.

The claim that process is the primary determinant of clock speed is easily dispelled with the Pentium 4 example.

 

[LightningZ71]

The 285K was selling for $420 at Microcenter a few weeks ago. That nuts for the top of the line part. I was tempted to buy one just to fool around with it.

It might be worth the adventure. Supposedly, with a LOT of overclocking work, you can get it up to reasonable gaming performance for the money in most titles.

 

[Khato]

If you're spending $1k+ for a graphics card then by all means, get the X3D processor to match it. Same if you care about performance on one of the games that disproportionately benefit from the increased cache. Otherwise? Good luck noticing a difference in processor performance when you're GPU bound.

 

[dullard]

If you have to delete the subject of a sentence to get it to say what you want it to say, you’re not reading it truthfully.

You do know that you are talking about a paraphrase of Gelsinger's actual comments--not his actual comments. And you are taking that one paraphrased sentence out of context. Gelsinger was asked about the Intel 3 and Intel 4 nodes. The reply mentioned Intel 3, Intel 4 and added this from Intel 18A:
"And then, of course, 18A as we deliver the PDK for that in Q2, the 1.0 PDK and we'll begin the volume ramps on Clearwater Forest, [Inaudible] Panther Lake in the first half of next year for those products coming out."

Intel (INTC) Q1 2024 Earnings Call Transcript | The Motley Fool

INTC earnings call for the period ending March 31, 2024.

www.fool.com

No where does that actually state Panther Lake is coming out mid-2025. The whole context of the entire question was about nodes. Gelsinger's actual statement was taking about 18A.

It is you who is deleting the subject.

 

[jpiniero]

No where does that actually state Panther Lake is coming out mid-2025.

He says later in the transcript they were planning on releasing Panther Lake in 2025 though. So ramping up production in mid 2025 for release at the end.

Obviously that isn't happening now.

 

[coercitiv]

But gaming is what drives DIY.

You just can't win with this crowd. I was making a point that ARL was in trouble at launch even when excluding gaming, leading to a very poor perception in DIY.

It might be worth the adventure. Supposedly, with a LOT of overclocking work, you can get it up to reasonable gaming performance for the money in most titles.

But the green monsta' does not play games. He does have a soft spot for lots of little cores doing video/image processing and other chores.

 

[Hitman928]

You do know that you are talking about a paraphrase of Gelsinger's actual comments--not his actual comments. And you are taking that one paraphrased sentence out of context. Gelsinger was asked about the Intel 3 and Intel 4 nodes. The reply mentioned Intel 3, Intel 4 and added this from Intel 18A:
"And then, of course, 18A as we deliver the PDK for that in Q2, the 1.0 PDK and we'll begin the volume ramps on Clearwater Forest, [Inaudible] Panther Lake in the first half of next year for those products coming out."

Intel (INTC) Q1 2024 Earnings Call Transcript | The Motley Fool

INTC earnings call for the period ending March 31, 2024.

www.fool.com

No where does that actually state Panther Lake is coming out mid-2025. The whole context of the entire question was about nodes. Gelsinger's actual statement was taking about 18A.

It is you who is deleting the subject.

Why use a transcript where they didn’t hear the comments clearly? I pulled mine where the full comment was captured. Additionally, it was part of his opening statement, not in response to any question. So yes, he did say it, you’re just looking in the wrong spot.

You can include the comment about the PDK if you want but it doesn’t change the fact that he starts a new sentence with a new subject for that sentence after the PDK comment.

Here’s a second source with the full transcript where the full comment is heard that says the same thing as my first source. This is a direct quote.

“We expect to release the 1.0 PDK for Intel 18A this quarter. Furthermore, our lead products, Clearwater Forest and Panther Lake are already in fab and we expect to begin production ramp of the Intel 18A in these products in the first-half of '25 for product release in the middle of next year.”

https://seekingalpha.com/article/4686485-intel-corporation-intc-q1-2024-earnings-call-transcript

 

[OneEng2]

It's a good product now, after numerous firmware updates and a considerable price drop.

When it launched, the 265K was just 2% faster than 14700K in Techpowerup's testing (performance in applications, gaming excluded). The worst offender in my book was the poor browser performance, making it slower than a 12700K in Speedometer or Jetstream. The pricing added insult to injury, as it matched the 9900X.

View attachment 132698

People bash Arrow Lake because it started on the wrong foot and Intel never did manage to shake this bad image.

One might say the same of Zen 5 as well. Intel just didn't seem to shake the initial bad feelings as quickly (or at all).

I also think that when you have the undeniable lead part (X3D) in the biggest DIY segment (gaming) it gives you quite a bit of slack in your downstream products with respect to perception.

The claim that process is the primary determinant of clock speed is easily dispelled with the Pentium 4 example.

Well.... that is a pretty extreme example. It was done at a time when processors were frequently limited in clock speed by pipeline misalignment while performing OOE.

Seems like today things are much more PPA limited .... which is much more a province of the process rather than the design (although that is not a hard and fast rule).

You just can't win with this crowd. I was making a point that ARL was in trouble at launch even when excluding gaming, leading to a very poor perception in DIY.

I noticed that! Tough crowd

 

[Doug S]

The reason I cheer on the E core team is because they are able to make big changes and new ideas while delivering. I know some will roll their eyes again, but it's plain simple. Look at the evolution of the E core. It's not following the big core. The amount of things they "break" and change is at multiples of what the P core did.

Yeah I agree that seems to be the team that is the best run, at least judging by results. The P core has been iterated and iterated over the years. Maybe they have done a true clean sheet in there a few times but if they have I can't point to a particular spot and say "that's where it happened", because it sure looks iterative to one degree or another all the way back to Pentium Pro. If anyone can absolutely confirm that there has been a clean sheet ground up redesign anywhere during that time I'd be interested in knowing when that happened and which CPUs would be an example of the last of the "old" design and the first of the "new" clean sheet design.

The E core team did start from a clean sheet and I think it shows based on their rapid iteration and improvement. The advantage of a clean sheet is that every part of the design has someone on the team who is fully responsible for it. Whereas if you have an old design you're tweaking to make wider, improve branch predictors, etc. there are parts of the design where the people who know it best are no longer on the team. The more of the core that's true for, the more difficult any major improvements become.

 

[regen1]

He says later in the transcript they were planning on releasing Panther Lake in 2025 though. So ramping up production in mid 2025 for release at the end.

Obviously that isn't happening now.

Many things(new CEOs, lay-offs, restructuring, etc.) have changed since. They have confirmed to launch a PTL SKU within 2025, if that happens, technically they got it down. CWF delay has been put down to issues with advanced packaging as per Intel.
Anyway for higher volumes on 18A, official position is still quite similar to what was back then.

"We will start to ramp to higher volumes in 2026 for 18A."

Intel's 18A-based CPUs won't ramp to high volume production until 2026

Intel will use more EUV in 2025

www.tomshardware.com

^April 2024

 

[Meteor Late]

I mean, if you downgrade 18A top speed, only achieving the same speed as Lunar Lake on already poor N3B process, then of course the delay won't be as big as it would've been otherwise.

 

[Khato]

CWF delay has been put down to issues with advanced packaging as per Intel.

Yeah, in typical Intel fashion design a product entirely reliant on the fancy new assembly technique. Seems like they've gotten it mostly figured out by now, but the delay hurts. Will be quite interesting to see how the end product stacks up given the design approach.

 

[Khato]

Seems like today things are much more PPA limited .... which is much more a province of the process rather than the design (although that is not a hard and fast rule).

I'd normally associate PPA with the design side again. Can easily increase frequency at the cost of area and power, either by addition of more pipeline stages or higher drive current transistors. We've already seen this done in some ARM designs, with one of the 'big' cores optimized for higher frequency than the rest at the cost of area and efficiency.

With 18A in particular there is something to be said for thermal density issues. I forget how long ago it was discussed previously, but it does sound like that's part of the issue with respect to max frequency on PTL. That's again something that design can take into account for future... and again a failing of the PDK. End result being that single thread performance may not improve, but efficiency and multi thread performance will.

 

[coercitiv]

One might say the same of Zen 5 as well.

I agree, Zen 5 felt underwhelming at launch on the consumer side, IIRC they had to drop prices until morale improved. There was a difference though, Zen 5 was more consistent: so it was seen as a very meh upgrade but an upgrade nonetheless.

I also think that when you have the undeniable lead part (X3D) in the biggest DIY segment (gaming) it gives you quite a bit of slack in your downstream products with respect to perception.

The Zen 5 X3D parts did fix public perception, and I agree it likely trickled downstream. Meanwhile Intel's fixes came in multiple waves of patches and it was really hard for consumers to gauge the improvements. My argument is ARL started on the wrong foot and never really recovered in terms of public perception (even though technically we know perf & value is fine now).

 

[DrMrLordX]

One might say the same of Zen 5 as well.

Arrow Lake didn't have an enterprise cousin that was a roaring success. Zen 5 had Turin and Turin dense which were smash hits.

 

[511]

I find it funny many times that 3 Intel Server Gen is three variants of Golden Cove

 

[DrMrLordX]

I find it funny many times that 3 Intel Server Gen is three variants of Golden Cove

If you can't laugh, you'll cry.

 

[511]

I do laugh on this part many times

 

[LightningZ71]

I find it funny many times that 3 Intel Server Gen is three variants of Golden Cove

Why? How many server gens did Xeon go where it was essentially just a VERY slightly tweaked version of the original SkyLake core with eventually a tacked on AVX512 unit? Kaby, Coffee and Comet Lake were similar scale or smaller changes, outside of the AVX512 unit, than what Golden Cove is getting along the way.

 

[ToTTenTranz]

Those points also apply to Lunarlake. Pantherlake uses the same Scalable Fabric Gen 2 as Lunarlake.

Unlike what we've seen for Panther Lake, I can't see any statement from Intel or tech journalists claiming the Xe2 and NPU in Lunar Lake can write into the L3. On the contrary, that feature in Panther Lake seems to be presented as a novelty.

 

[Henry swagger]

I agree, Zen 5 felt underwhelming at launch on the consumer side, IIRC they had to drop prices until morale improved. There was a difference though, Zen 5 was more consistent: so it was seen as a very meh upgrade but an upgrade nonetheless.

View attachment 132708


The Zen 5 X3D parts did fix public perception, and I agree it likely trickled downstream. Meanwhile Intel's fixes came in multiple waves of patches and it was really hard for consumers to gauge the improvements. My argument is ARL started on the wrong foot and never really recovered in terms of public perception (even though technically we know perf & value is fine now).

I agree .. tsmc n3b wasnt impressive for using euv vs intel 7