Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads

Page 19 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Tigerick

Senior member
Apr 1, 2022
820
785
106
PPT1.jpg
PPT2.jpg
PPT3.jpg



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.



LNL-MX.png

Intel Core Ultra 100 - Meteor Lake

INTEL-CORE-100-ULTRA-METEOR-LAKE-OFFCIAL-SLIDE-2.jpg

As mentioned by Tomshardware, TSMC will manufacture the I/O, SoC, and GPU tiles. That means Intel will manufacture only the CPU and Foveros tiles. (Notably, Intel calls the I/O tile an 'I/O Expander,' hence the IOE moniker.)



Clockspeed.png
 

Attachments

  • PantherLake.png
    PantherLake.png
    283.5 KB · Views: 24,028
  • LNL.png
    LNL.png
    881.8 KB · Views: 25,521
Last edited:

jpiniero

Lifer
Oct 1, 2010
16,713
7,176
136
It has to be More Huge Jump of IPC(ISO Speed) because it has to make up for the 33% core deficit on the P core side.

Raptor Lake 8+16 vs MTL 6+16

That's what makes that rumor dumb. Reusing the 6+8 P CPU die would make some sense, doing a seperate 8+16 die (if by some miracle yields allow for it) would make sense. But not 6+16.
 

mikk

Diamond Member
May 15, 2012
4,293
2,382
136
It has to be More Huge Jump of IPC(ISO Speed) because it has to make up for the 33% core deficit on the P core side.

Raptor Lake 8+16 vs MTL 6+16

MTL-S and ARL-S are reportedly both branded as 14th gen, MTL-S will serve for the low-midrange and ARL-S for the higher end (likely limited to Core i7-Core i9). They obviously won't replace MTL-S 6+16 with RPT 8+16 but they can easily replace the current RPT/ADL i5 lineup with additional +4 E-cores.
 

jpiniero

Lifer
Oct 1, 2010
16,713
7,176
136
MTL-S and ARL-S are reportedly both branded as 14th gen, MTL-S will serve for the low-midrange and ARL-S for the higher end (likely limited to Core i7-Core i9).

It's sounding more like Arrow is going to be 15th Gen. They can always fill in the Meteor gaps by doing Raptor rebrands.
 

A///

Diamond Member
Feb 24, 2017
4,351
3,160
136
a very old leak pointed at arrow lake s getting 8+32 from a now removed benchmark that came online. This latest rumor is strange because if you were to go with go with what @mikk said you're looking at the i7 to i9 having 3 skus or that's the i5, i7 and i9 having the same core setup but with different top clock and all core, or it's i7, i9 and i9 ks edition.

I'd like to see more but am surprised intel is sticking with the current layout available on raptorlake. This is either bs and fabricated or intel has managed to design their p and e cores to get monster performance out of them on a multi die setup and they're confident it'll out phase whatever amd has in store for then.


Mlid wasn't the first person to report the 8 + 32 setup for arrow, he was very late on that. apart from him being good marketing material on the safety practices and failure rates of latex products used in an intimate setting he should be ignored.
 

dangerman1337

Senior member
Sep 16, 2010
384
45
91
That's what makes that rumor dumb. Reusing the 6+8 P CPU die would make some sense, doing a seperate 8+16 die (if by some miracle yields allow for it) would make sense. But not 6+16.
Yeah, reusing the MTL-P for i5 14600K makes sense but 6+16 chiplet for an i7, i9 etc is bizzare.

Like why? I wonder if Raichu on Twitter can give us an answer.
 

mikk

Diamond Member
May 15, 2012
4,293
2,382
136
a very old leak pointed at arrow lake s getting 8+32 from a now removed benchmark that came online. This latest rumor is strange because if you were to go with go with what @mikk said you're looking at the i7 to i9 having 3 skus or that's the i5, i7 and i9 having the same core setup but with different top clock and all core, or it's i7, i9 and i9 ks edition.

I'd like to see more but am surprised intel is sticking with the current layout available on raptorlake. This is either bs and fabricated or intel has managed to design their p and e cores to get monster performance out of them on a multi die setup and they're confident it'll out phase whatever amd has in store for then.


Mlid wasn't the first person to report the 8 + 32 setup for arrow, he was very late on that. apart from him being good marketing material on the safety practices and failure rates of latex products used in an intimate setting he should be ignored.


The 8+32 leak came from the reddit leak which is quite old, who knows what changed since then or who knows if it was correct. The new SKU infos apparently comes from an Intel slide which is much more reliable. Both MTL-S and ARL-S seems listed in the same document which also suggest they are launching it together under the same 14th Gen brand.
 
  • Like
Reactions: Tlh97 and A///

A///

Diamond Member
Feb 24, 2017
4,351
3,160
136
The 8+32 leak came from the reddit leak which is quite old, who knows what changed since then or who knows if it was correct. The new SKU infos apparently comes from an Intel slide which is much more reliable. Both MTL-S and ARL-S seems listed in the same document which also suggest they are launching it together under the same 14th Gen brand.
can you link me? I don't visit that hellhole. if this slide leak is true does that not mean that arl-s is not being made at tsmc? arl-s was reportedly coming out in late 2024 but this new leak would place it at end of q3 in 2023? I'm not doubting intel if they have it set in stone compared to their dc stuff. could this be the gelsinger magic everyone's been reporting about ever since bird brain swan was cast to the side like a roast goose?
 

inf64

Diamond Member
Mar 11, 2011
3,884
4,692
136
So Meteor Lake and Arrow Lake will have different P cores and are supposed to launch at the same time? if that's the case it makes little sense to me (or I misunderstood something).
 

A///

Diamond Member
Feb 24, 2017
4,351
3,160
136
So Meteor Lake and Arrow Lake will have different P cores and are supposed to launch at the same time? if that's the case it makes little sense to me (or I misunderstood something).
two process lines, one an older node, one leading edge. think of rpl's lower end chips being tarted up alderlake as a foot in the door for intel to do this. it makes sense. Lower class chips don't need leading edge performance on a newer node. We've seen this with how intel stretched 14nm and now 10 nm. for mtl as the low to mid end it makes sense to use an older node to save more money on the bom than a leading edge node like for arrow lake. it also frees up capacity for arrow and xeon if they use that same node.
 

A///

Diamond Member
Feb 24, 2017
4,351
3,160
136
if 8+16 is confirmed later on through leaks then intel is not playing the e core spam game as we thought. AMD to my knowledge does not intend to use e cores when zen 5 is a whole new uarch. intel p and e cores are changing for 14 and 15th now 14th as a whole generation. they must have uber confidence in them cores.
 

Geddagod

Golden Member
Dec 28, 2021
1,513
1,613
106
Why does Redwood Cove still include AVX512 when it's already confirmed that that will be disabled again? Is the core only a straight port of Golden Cove onto a new node?
I'm guessing so, maybe with some slight changes
But I also think it's likely that Redwood Cove (with a slight upgrade as Pat said) would be used in Granite Rapids.
 

mikk

Diamond Member
May 15, 2012
4,293
2,382
136
can you link me? I don't visit that hellhole. if this slide leak is true does that not mean that arl-s is not being made at tsmc? arl-s was reportedly coming out in late 2024 but this new leak would place it at end of q3 in 2023? I'm not doubting intel if they have it set in stone compared to their dc stuff. could this be the gelsinger magic everyone's been reporting about ever since bird brain swan was cast to the side like a roast goose?


Link for what? I don't understand your TSMC question to be honest. Regardless of the core count we cannot confirm or deny any process node rumors. With TSMC 3nm they can launch both at the same time, to me it makes sense. If ARL-S comes on 20A it won't be ready before H2 2024 realistically. In this case they can launch LGA1851 limited to Core i5 level of performance in H1 2024 or they have to delay MTL-S into H2 2024 to bring both at the same time, both options are not so great. And MTL-S in 2023 is not realistic, Intels primary focus clearly is on MTL-M/P in 2023.
 

A///

Diamond Member
Feb 24, 2017
4,351
3,160
136
I'm guessing so, maybe with some slight changes
But I also think it's likely that Redwood Cove (with a slight upgrade as Pat said) would be used in Granite Rapids.
Is there a chance they figured out and helped microsoft design thread director to work with avx512? afaik thread director is in its 2nd gen with raptor lake and windows 11. Which is still a problem because you have a great processor coupled with a crippled hobbled together half dead hobo of a os.
 

Exist50

Platinum Member
Aug 18, 2016
2,452
3,106
136
Did you expect 6+8?
Yeah, I said that the time that that rumor sounded correct. 6+16 is new to me, but would also make sense. Either way, I don't think they intend Meteor Lake as a successor to Raptor Lake across the complete lineup.

I think it's also possible, even likely, that Intel's been adjusting core counts for the 2024 products. Should be pretty simple for them to cut or add to the existing compute dies. But to lower the top end ARL while increasing the top end MTL is certainly interesting. Volume, margin, performance issues? No idea.
It has to be More Huge Jump of IPC(ISO Speed) because it has to make up for the 33% core deficit on the P core side.
We're not getting a "more huge jump" in IPC out of MTL, so two possibilities. 1) The iso-power gains from Intel 4 and the new archs are enough to beat RPL despite -2 P cores. OR 2) They don't intend for it to do so.
 
  • Like
Reactions: Tlh97 and coercitiv

Exist50

Platinum Member
Aug 18, 2016
2,452
3,106
136
can you link me?
The original post was deleted, but someone grabbed a copy of it:
for mtl as the low to mid end it makes sense to use an older node to save more money on the bom than a leading edge node like for arrow lake.
If it were just a matter of different nodes, I think they'd much prefer to have LNC+SKT across the lineup. My feeling is that MTL delays have kinda compressed the two together in a way that'll make for a messy product line.
 

Exist50

Platinum Member
Aug 18, 2016
2,452
3,106
136
Not that Huge

Golden Cove on the left and Redwood Cove on the right
View attachment 70090
Tbh, that seems pretty darn large for what's supposedly a node with 2x the density.
Why does Redwood Cove still include AVX512 when it's already confirmed that that will be disabled again? Is the core only a straight port of Golden Cove onto a new node?
Seems like it, or near enough, though it's probably also worth taking Locuza's annotations with a grain of salt.
 

poke01

Diamond Member
Mar 8, 2022
4,067
5,391
106
If Lion Cove is not Royal and Arrow Lake is supposed to big. So Royal is even more impressive than Lion?

so it would be Royal > Lion > Redwood
 

msj10

Member
Jun 9, 2020
92
113
106
If Lion Cove is not Royal and Arrow Lake is supposed to big. So Royal is even more impressive than Lion?

so it would be Royal > Lion > Redwood
There's also Panther cove after Lion so it seems Royal is at least 3 generations away