Discussion Intel current and future Lakes & Rapids thread

[dullard]

Ok, last one for today. This isn't a guarantee that Meteor Lake is coming out on desktop, but at least it is something that Intel has considered: https://www.intel.com/content/www/us/en/search.html?ws=related#q=meteor lake_s&sort=@createddate descending
Click "Show sub-topics" and you see:

Enabled Simics® simulator support for Arrow Lake S CPU / Meteor Lake PCH-S.

Using Intel's terminology PCH-S is the platform controller hub for desktop chips: https://edc.intel.com/content/www/u...form-controller-hub-pch-datasheet-volume/003/

PCH-S refers to desktop and server/workstation PCH SKUs

 

[Ajay]

Understatement of the week. 104 cores of ADL scoring 80K in CB23 at 2x370W power taken as gospel is hilarious.

Using CB23 to compare server CPUs is also hilarious. Not pointing this at you, but we keep posting this nonsense wrt server CPUs year after year, even though it has almost no bearing on any actual workloads modern servers are running.

 

[Exist50]

Frankly, I'll be very surprised if they ship MTL at the range they detailed on that slide. Seems more likely that it'll only go down to ~9W and up to ~95.

 

[JoeRambo]

Using CB23 to compare server CPUs is also hilarious. Not pointing this at you, but we keep posting this nonsense wrt server CPUs year after year, even though it has almost no bearing on any actual workloads modern servers are running.

Yeah, 100% agreed, in fact CB has nothing to do with desktop and laptop PCs either. Only small fraction is running rendering loads ( could be even smaller than amount of Xeons pressed in workstation duty ).
But CB23 is a funny load that does not care about memory subsystem that much as long as L2 fits it, and has "expected" scores of where each core should stand. So in this case CB23 provided a good reality check of where 104 cores of ADL with 2x370W power budget should score. So armed with this reality check, we know that other MT scores are shady and whole MT comparison is bogus as well.

 

[DrMrLordX]

How can that be the first you are hearing that?

You're focusing on the wrong thing. It was the first time I've heard that Meteor Lake was coming to desktop (or back to desktop, depending on how you look at it). I have heard about Arrow Lake on 20a though I'm still skeptical about that happening.

 

[dullard]

You're focusing on the wrong thing. It was the first time I've heard that Meteor Lake was coming to desktop (or back to desktop, depending on how you look at it). I have heard about Arrow Lake on 20a though I'm still skeptical about that happening.

That is what is confusing to me. There have been rumors for years about Meteor Lake desktop. That is why I posted many links about those rumors (I posted just a small subset of the links since many are duplicated across all the rumor websites). Desktop Meteor Lake was even discussed in this thread. So, I am confused as to how it can be the first you are hearing about it.

What would you like me to focus on?

 

[nicalandia]

Understatement of the week. 104 cores of ADL scoring 80K in CB23 at 2x370W power taken as gospel is hilarious.

Pretty much. Basic math using CB23 performance data for Golden Cove says an 80K score is what a single SPR 52 core SKU @ 370W should be able to achieve.

On What World would a 2 Ghz Golden Cove CPU that has 52 cores be able to achieve 80K? I will be generous with the Golden Cove and assume that CineBench R23 would scale 100%(it will not). You will get 41,600 points on a perfect world or about 38,000 points on a real world. Good Luck keeping a 52 Core at 4Ghz All Cores under 700 Watts, because that's exactly how fast those need to be to reach 80K points in CBR23

A single core at 5.2 Ghz gets about 2,000 points. Do your Basic Math PLEASE..

 

[JoeRambo]

You will get 41,600 points on a perfect world or about 38,000 points on a real world.
Do your Basic Math PLEASE..

So basically after Your "math", Intel is better off selling 2 socket 12900K @ 5Ghz to score same 40k, and achieve with 16C at 5ghz what 52C do at 2ghz ?

Hint, reality checks start with:

1) Core count difference is 3.25
2) Clock difference is 2.5

So 30% missing already in Your "math". Then there question of powers, 2x12900 @5ghz running cb23 would be say 400W, and 52C @ 2Ghz would be 370W? I doubt it needs to stay at 2Ghz having 370W of TDP, but obviuosly impossible to say what clocks final chips will run under CB23.

So we have 40k * 1.3 * unknown clock factor. And over two sockets this does not come out to what was claimed as score for 104 cores in the article.

 

[coercitiv]

Do your Basic Math PLEASE..

I did my math mate, I actually did TESTING before the math. ADL Golden Cove core with ~7.5W budget sustains ~3.4Ghz in CB23. With 62W package power 8c/16t Golden Cove scored 14K in CB23. Lower that score to around 12K+ for under 7W per core, multiply by 6.5X for a 52 core SKU and you get ~80K for a single CPU.

So get your facts straight and please stop echoing stupid press articles that are meant for clicks only. There's a reason why people who are trully interested in leaks thread carefully when being presented with raw data without clear details of how it was obtained.

 

[nicalandia]

I did my math mate, I actually did TESTING before the math. ADL Golden Cove core with ~7.5W budget sustains ~3.4Ghz in CB23. With 62W package power 8c/16t Golden Cove scored 14K in CB23. Lower that score to around 12K+ for under 7W per core, multiply by 6.5X for a 52 core SKU and you get ~80K for a single CPU.

You can't be this serious. The ES CPU is at 1.9 Ghz Not at 3.4 Ghz. And you can't just say. Hey my Desktop Alder Lake CPU(P core) is only using 7 Watts on single core and then multiply it by 52. It does not work like that. Have you seen how much more Xeons consume when compared to desktop parts due to fabric interconnect(Mesh vs Ring Bus, Alder lake can turn off the ring buss from the e cores when not used)

 

[nicalandia]

So basically after Your "math", Intel is better off selling 2 socket 12900K @ 5Ghz to score same 40k, and achieve with 16C at 5ghz what 52C do at 2ghz ?

Hint, reality checks start with:

1) Core count difference is 3.25
2) Clock difference is 2.5

So 30% missing already in Your "math".

Alder Lake is 8C/16T of P cores and 8C/8T of E cores so total would be 24 Threads

But, a 8C/16TGolden Cove at 5 Ghz all cores get about 20,000 points so a 16C/32T at half the speed should get about the same.

Edit.
need to take my morning coffee

 

[JoeRambo]

Alder Lake is 8C/16T of P cores and 8C/8T of E cores so total would be 24 Threads

And what default configuration has to do with the fact that 8C16T of ADL scores ~20K of CB23 @ 5Ghz ? Anyway i am done with this basic math checking disguising as trolling.

 

[nicalandia]

And what default configuration has to do with the fact that 8C16T of ADL scores ~20K of CB23 @ 5Ghz ? Anyway i am done with this basic math checking disguising as trolling.

Dude cut me some slack. I am coming of a 24 shift here. I had things confused. I got the single core and messed from that point..

But the mesh interconnect is going to become a real hassle for Sapphire Rapids as it did with Xeons

 

[igor_kavinski]

I am coming of a 24 shift here.

GO. TO. SLEEP. NOW!

 

[coercitiv]

You can't be this serious. The ES CPU is at 1.9 Ghz Not at 3.4 Ghz.

Here's ADL-S 8c/16t @ 2GHz. Notice the 26W package power?

And you can't just say.Hey my Desktop Alder Lake CPU(P core) is only using 7 Watts on single core and then multiply it by 52. It does not work like that. Have you seen how much more Xeons consume when compared to desktop parts due to fabric interconnect(Mesh vs Ring Bus, Alder lake can turn off the ring buss from the e cores when not used)

Watch me saying it. By ignoring uncore contribution on ADL-S I also ended up multiplying uncore power on desktop Golden COve by 6.5X times when normalizing for 52 cores. Even then I intentionally lowered score expectation for 8 cores from 14K to 12K just to account for even more uncore power. That's a ~15% lower score just to make sure the mesh and the tile design has room even after the 6.5X uncore multiplication from the desktop chip.

Now go back to the screenshot posted above, 26W for 8c/16t @ 2Ghz. You know what that means for 52 cores before the uncore? Less than 160W for 52 Golden Cove cores. Sleep on that.

 

[JujuFish]

Dude cut me some slack.

This coming from the person who posted "Do your Basic Math PLEASE". Pretty rich.

 

[DrMrLordX]

That is what is confusing to me. There have been rumors for years about Meteor Lake desktop.

Those were put to rest, quite some time ago. And those were just rumours. It was never anything official.

 

[dullard]

Those were put to rest, quite some time ago. And those were just rumours. It was never anything official.

Ok, now that makes sense. You had heard of desktop Meteor Lake, but accepted the rumors that it might not exist.

You are correct, no rumors for or against a new product are ever official until the product is announced. Any product can be canned at any point in the development. Heck, products can be cancelled even after announcement to the humiliation of the company. You won't find a single link from Intel proving that Meteor Lake desktop will launch. But you also won't find a single link from Intel that Meteor Lake desktop will not launch.

All we have is speculation from rumors and mounting evidence will support some rumors and refute others. It makes these discussions fun. I happen to like the little hints that we can find here and there to put the puzzles together. Here is the last puzzle piece that I know of right now (and I admit that it is a weak piece of evidence):

Power Supply Selector

Choose the proper power supply form factor and power level for your application.

www.intel.com

A search for Meteor Lake on Intel's website leads to a link for desktop power supply designs. Of course, that could be a mistake by Intel or they could change their minds up until the day it is shipped out. But that weak piece of evidence, along with the repeated Intel 125 W TDP announcements, and the Meteor Lake-S reference I linked above all point towards the existence of at least a prototype desktop Meteor Lake chip. That might not make it to market, but the evidence points towards it happening.

 

[DrMrLordX]

Ok, now that makes sense. You had heard of desktop Meteor Lake, but accepted the rumors that it might not exist.

On the contrary, the only rumour was that it did exist.

 

[Exist50]

On the contrary, the only rumour was that it did exist.

There were rumors that MTL was mobile only. Not backed by anything concrete, but there were indeed rumors.

And I don't think it's an entirely unjustified position. Anyone expecting a MTL release to mirror ADL will probably be sorely disappointed.

 

[dullard]

On the contrary, the only rumour was that it did exist.

Um what do we have here? https://www.hardwaretimes.com/intel...oks-arrow-lake-w-40-cores-for-desktops-rumor/

The latest whispers from the industry indicate that Intel will be skipping the Meteor Lake lineup on the desktop platform, launching Arrow Lake instead.

Is this also not a rumor?

 

[eek2121]

Unless the CPU chiplet is being fabbed at TSMC there is no way there will be a desktop version. They can't hide the volume like they can on mobile.

Meteor Lake (Intel 4) only uses a bit of EUV. Much of it will be on the same type of equipment they are already using for Alder Lake. It is Intel 3 with much more EUV needed where they really don't have the capacity to fab enough.

Here are some desktop Meteor Lake rumors:

• Mar 23, 2021: https://videocardz.com/newz/intels-first-7nm-client-architecture-meteor-lake-to-launch-in-2023 "Meteor Lake is Intel’s first 7nm client architecture, which means it will launch for desktop and mobile users under the Intel Core series." Note: Intel has since switched this naming to Intel 4 instead of 7 nm.
• May 24, 2021: https://videocardz.com/newz/intels-7nm-meteor-lake-compute-tile-has-taped-in "The series is rumored to be the last to support the LGA1700 socket" This LGA1700 socket is a desktop socket.
• May 26, 2021: https://www.techadvisor.com/news/pc-components/intel-meteor-lake-14th-gen-cpu-3803533/ "Will Intel release Meteor Lake desktop and mobile chips? Most likely, although their releases may be staggered".
• July 26, 2021: https://videocardz.com/newz/intel-r...re-rapids-meteor-lake-and-granite-rapids-cpus "Intel confirmed that Meteor Lake will use Foveros packaging technology and it will support TDP ranges from 5 to 125W."
  • This 125 W is direct from Intel itself, slide 35: https://download.intel.com/newsroom/2021/client-computing/Intel-Accelerated-2021-presentation.pdf There are not that many 125 W CPUs that are mobile only.
• Nov 18, 2021: https://videocardz.com/newz/intel-14th-gen-core-meteor-lake-mobile-test-cpu-has-been-pictured "Intel Meteor Lake desktop and mobile CPU series should not be expected till the second quarter of 2023."
• Feb 18, 2022: https://wccftech.com/intel-next-gen...gen-arrow-lake-2024-16th-gen-lunar-lake-2025/ "Intel Confirms Raptor Lake, Meteor Lake, Arrow Lake, Lunar Lake Client Desktop CPU Platforms"
• Apr 28, 2022: https://videocardz.com/newz/14th-gen-core-meteor-lake-has-been-powered-on-on-track-to-launch-in-2023 "Therefore, Meteor Lake will debut on low-power laptops as well as high-end desktop consumer systems."

Of course those are mostly rumors or forward-looking statements from Intel. Thus things in the future can change. But for that to be DrMrLordX's first time hearing of desktop Meteor Lake meant his eyes were closed to the rumors.

Maybe it's a limited desktop launch (in 2023). Like ADL-S being K-SKU limited in 2021. A full scale desktop lineup is unlikely because of the volume (EUV) limitation in 2023.




Depends on the source, most of your links are irrelevant, they have no source. There are only two reliable sources for a desktop version as far as I know, the one where Intel told us it will scale from 5-125W and the second is MLID, he claimed there will be a desktop version.

Frankly, I'll be very surprised if they ship MTL at the range they detailed on that slide. Seems more likely that it'll only go down to ~9W and up to ~95.

I suspect all the rumors that some here keep repeating stem from the (false) rumor that Intel has an EUV shortage. Intel published their EUV numbers during investor day. They have far more EUV capacity than will ever be needed short and medium term. Long term they are working with ASML to aggressively scale up High-NA EUV. Sure, it's okay to be skeptical , but that doesn't mean that they aren't better set up than most here will ever admit.

 

[DrMrLordX]

Is this also not a rumor?

No, considering the amount of supporting information that showed laptop-only Meteor Lake, initially for 2022 on Intel 7nm/Intel 4. Before that it was an unknown entity (e.g. no indication of whether it would be mobile or desktop), associated only with Ocean Cove. And by that I meant the first Ocean Cove, not the second one that we'll actually see in 2023.

If you wish to regard it as a rumour then by all means, go ahead.

To return to my initial point, up to this time, I have heard no one definitively state that Meteor Lake would, in fact, be a desktop product. Initially it was an unknown entity, associated with Ocean Cove (the first one, not the second one). Then it was placed solidly as a mobile product, and we were even shown some mockups of the tile layout which seemed consistent with the idea that it would be for laptops and not desktops. And now we're hearing it may be for desktop as well. I guess that makes sense since without Arrow Lake in 2023, Intel will not have a desktop product after Raptor Lake until 2024.

 

[dullard]

No, considering the amount of supporting information that showed laptop-only Meteor Lake, initially for 2022 on Intel 7nm/Intel 4. Before that it was an unknown entity (e.g. no indication of whether it would be mobile or desktop), associated only with Ocean Cove. And by that I meant the first Ocean Cove, not the second one that we'll actually see in 2023.

If you wish to regard it as a rumour then by all means, go ahead.

To return to my initial point, up to this time, I have heard no one definitively state that Meteor Lake would, in fact, be a desktop product. Initially it was an unknown entity, associated with Ocean Cove (the first one, not the second one). Then it was placed solidly as a mobile product, and we were even shown some mockups of the tile layout which seemed consistent with the idea that it would be for laptops and not desktops. And now we're hearing it may be for desktop as well. I guess that makes sense since without Arrow Lake in 2023, Intel will not have a desktop product after Raptor Lake until 2024.

Please show that "supporting information" that placed meteor lake "solidly as a mobile product". That would help the discussion. For desktop Meteor Lake, the most definitive evidence is still this Intel image:


I don't personally think desktop Meteor Lake will be around for long. It'll be more of an afterthought with Intel 3 coming along quite quickly after it is launched.

 

[DrMrLordX]

Please show that "supporting information" that placed meteor lake "solidly as a mobile product".

It's been littered through this thread and elsewhere in the forum. The same one we read. How long you got? I don't currently have the time to search that up but if you really need reminders I can hunt it all down. Meanwhile:

https://www.tomshardware.com/news/intel-meteor-lake-pictured

That's one of the mockups that Intel has been showing off. Not looking like a desktop chip. The core count looks too low.