Discussion Intel current and future Lakes & Rapids thread

Page 725 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
Can we go with the date that reviews are available and at least some people in multiple countries have purchased it successfully?
I would say Reviews in the western world as someone in China may get a hold of a laptop sample and review it many months before it reaches a USA/UK/Euro based reviewer.
 

Geddagod

Golden Member
Dec 28, 2021
1,147
1,003
106
No that's not him. I forgot about bepo, but ye I think he was the first person to say Intel 4 was going to clock bad.
Intel did share that graph btw, and it showed Intel 4 better throughout the power curve vs Intel 7 an reaching higher clocks, but some people are concerned about the max frequency of Intel 4.
 

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
Nah, as @Geddagod said, it was david bepo on twitter who's been claiming huge, GHz+ frequency penalties for Intel 4. Can't find the precise tweet in question, but it's not like the specifics matter. Though MLID has also claimed an unspecified penalty.

Though obviously it's hard to get a truly apples to apples comparison. Even if RWC is basically just a GLC shrink, backend optimizations can swing things by a couple hundred MHz in some cases. I wouldn't be surprised if MTL misses RPL clocks, but I think the processes should be at least roughly comparable at peak.
 

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
I will be shocked if we don’t see a new product out of Intel next year. If the rumors of multiple X3D AMD SKUs are true, Intel is going to struggle with sales.

I think Meteor Lake will land in q3 or q4 of next year.
 

lightisgood

Member
May 27, 2022
156
64
61
*****machine translation*****
Mass production ahead of schedule in 2024 Intel's big move "1.8nm" is not hidden: the list of customers will be announced next year
> Intel has not mentioned who the specific customers of the 18A process are, which is also very concerned by the market,
> and now the Intel CFO has finally given a time point, saying that they will publish the list of 18A customers early next year,
> and it seems that there is more than one, and there is even a company closely related to the US Department of Defense.

If Intel 4 is never, Intel 3 is never, Intel 20A is never... BUT Intel 18A is sure in 2H24?
I suspect Intel 18A has been proved to be financial by non-competitive contract.
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
I will be shocked if we don’t see a new product out of Intel next year.
Yes, there will be plenty of new products, for starter the non K SKUs(Alder Lake Refresh), the H, HS, HX Raptor Lake models and of course the KS model at the very top 13900KS.

For 2024 they will likely do a Raptor Lake Refresh with Intel 7+ and different SKUs.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Does anyone think the rumored Raptor Lake refresh could be an Intel 4 die shrink?

No.

I will be shocked if we don’t see a new product out of Intel next year.

Does Raptor Lake refresh count? That's Intel's chance to put more of their 13-series products on actual Raptor Lake, rather than the small number of SKUs available now.

If the rumors of multiple X3D AMD SKUs are true, Intel is going to struggle with sales.

The performance difference will likely be ugly, but we don't know what the volume will be like on those X3D parts.

I think Meteor Lake will land in q3 or q4 of next year.

My prediction is limited volume in Q4, with full volume by Q1 of 2024.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
That might not be anything to you, but in the big picture that is a pretty significant benefit.

Tell me you are kidding.

Technology is not the solution, people are. Uncontrolled usage is the problem. If EVERYONE was careful of spending(money/power/resources) then it'd be 100x any tech would do you. Also that is parroting what the elites are trying to tell you, that the lowly citizen are at fault, when big portion of the problem is THEIR policy.
 
Last edited:

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
I still stick to the idea that July-September 2013 we'll see 5W Meteorlake, since the Jan 3 launch doesn't include Raptorlake M. This is very similar to the Core M situation. Then a few months later(likely Jan 3 2024) we'll see Meteorlake-S/U/P/H. This also mirrors the situation with 14nm, where it had a difficult ramp.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
The process might be, but that doesn't mean the design/architecture is ready.

This timeframe was much closer before. It was merely a quarter gap. If the improve execution, they might close it again. But yea CPUs are more than just process being ready now. Chiplets complicate this equation further.
 

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
If we're all spitballing, might as well throw my hat into the ring. I'd say 50% confidence interval for an announcement/launch event in October-ish, with a small selection of premium devices on shelves for Black Friday/Holiday shopping. 90% confidence interval would be the same criteria, but announcement delayed till CES, so basically +1Q. Either way, I'd expect more mainstream laptop availability to lag roughly a quarter behind the premium/flagship ones.

Think they'll start with U/P, and announce M and H separately sometime in the following months. S, rumors seem to say is coming quite a bit later.
 

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
I still stick to the idea that July-September 2013 we'll see 5W Meteorlake, since the Jan 3 launch doesn't include Raptorlake M. This is very similar to the Core M situation. Then a few months later(likely Jan 3 2024) we'll see Meteorlake-S/U/P/H. This also mirrors the situation with 14nm, where it had a difficult ramp.

There really isn't an M anymore and Raptor-U is included in the Jan 3 rumors. If there's no "9" W versions though, you might be on to something.

I doubt you will see Meteor-S actually get released given the Raptor Refresh rumors. You can do both on mobile but not on desktop. I suppose they could do 65 W B models like they did with Tiger Lake but I don't think you can really call that desktop.
 

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
I doubt you will see Meteor-S actually get released given the Raptor Refresh rumors.
I think they'll still do a desktop release of some kind for the same reasons that pushed them to release Rocket Lake. But hopefully they've learned not to push that into the enthusiast/gaming market.
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
Raptor Lake-S Refresh leaked as true. Meteor Lake-S not showing up, at best, until early 2024.
That is true, MTL-S is not a 2023 product. I don't think we will ever see a MTL-S. Just the mobile Limited SKUs by late 2023 in China.
 
  • Like
Reactions: ftt

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
Last edited:
  • Like
Reactions: igor_kavinski

Geddagod

Golden Member
Dec 28, 2021
1,147
1,003
106
RPL-R speculation
High probability:
+100-200 MHz clocks (6Ghz)
improved ringbus speeds
Maybe:
+300 MHz clocks (6GHz+)
increased L3 cache (rumored to be in Emerald rapids)
negligible to slight IPC uplifts (from cache and maybe some more tweaks)
This is just for fun :)
architecture is a backport of meteor lake's redwood cove (maybe a 5 or so percent uplift in IPC)
separate L4 cache like on Broadwell
Trolling:
RPL-R is a monolithic CPU on Intel 4 (LOL)
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
RPL-R speculation
High probability:
+100-200 MHz clocks (6Ghz)
improved ringbus speeds
Maybe:
+300 MHz clocks (6GHz+)
increased L3 cache (rumored to be in Emerald rapids)
negligible to slight IPC uplifts (from cache and maybe some more tweaks)
This is just for fun :)
architecture is a backport of meteor lake's redwood cove (maybe a 5 or so percent uplift in IPC)
separate L4 cache like on Broadwell
Trolling:
RPL-R is a monolithic CPU on Intel 4 (LOL)
You are being too optimistic on your assumptions. RPL-R is not a backport of MTL.

Emerald Rapids is just Raptor Lake on the server.
 

Geddagod

Golden Member
Dec 28, 2021
1,147
1,003
106
You are being too optimistic on your assumptions. RPL-R is not a backport of MTL.

Emerald Rapids is just Raptor Lake on the server.
Which is why I headed the RPL-R being a backport with "this is just for fun"
Raichu thinks EMR has increased L3 cache compared to SPR. Or maybe L2, he just said "increased cache in the interconnect".
Also Raptor Cove seems to be closer to GLC server than anything else really. Raptor Cove and GLC server have the same amount of L2 cache at 2MB. The only improvements RPL seems to bring is a slightly tweaked prefetcher and a slightly tweaked node.
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
I dare say that the W7-2495X with 24 P cores, 48 threads and 4 full DDR5 channels is going to be universally faster than the 13900KS in MT and close enough in ST situations there it won't matter.

1670511233096.png

4 Memory Channel, 45 MiB of total L3 is not gonna cut it for HEDT anymore.


Can we just agree that Intel HEDT-X line is just not going to happen and that they are just releasing Xeon Workstation product SKUs for the 1S segment and their 2S Workstation System will be just Regular Xeon Platinum based on Sapphire Rapids.

How many of you are impressed by The Xeon W SKUs?
 
Last edited:
  • Like
Reactions: lightmanek

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
View attachment 72497

4 Memory Channel, 45 MiB of total L3 is not gonna cut it for HEDT anymore.

There is a 3000 series that has the full 8 channels. This is probally the "LCC" - 24 cores and half the memory channels and lanes. It's still going to be a very big die for 10 nm. Whether these products make sense to buy versus say Threadripper Pro is an open question sure.

Edit: And pretty sure Xeon W has always been 1S only.
 
Last edited:

dullard

Elite Member
May 21, 2001
24,998
3,325
126
Can we just agree that Intel HEDT-X line is just not going to happen
So you post a picture listing four -X processors and post that -X is not going to happen in the same post? Please expand on what it is you are trying to say. Otherwise what you posted looks like gibberish. It seems like you are trying to add opinions as fact without stating your opinions.