• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Discussion Intel current and future Lakes & Rapids thread

Page 442 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

jpiniero

Lifer
Oct 1, 2010
10,085
2,349
136
6+0 isn't usable for cheap OEM devices, no iGPU.
The 6+0 does have an IGP, or at least I would be surprised if it doesn't. Just no small cores.

Guessing they aren't confident enough in Alder Lake supply and want to ensure that the low end is covered with something. This isn't a problem with laptops since everything is BGA.
 
  • Like
Reactions: Tlh97 and mikk

dullard

Elite Member
May 21, 2001
23,037
1,294
126

mikk

Diamond Member
May 15, 2012
3,242
1,057
136
Icelake taped in around June 2017 and it came out in August 2019. Without the big 10nm struggle we would haven seen Icelake early 2019. Unless there are another big 7nm delays we should get MTL in H1 2023. And Intel not long ago revealed it's a H1 2023 platform , so this makes sense. Unlike Videocardz I believe MTL will be mobile only anyways, it's a new process and from 14nm Broadwell-U and 10nm Icelake-U we do know Intel prefers mobile first. First 7nm desktop is more likely Lunar Lake.
 

uzzi38

Golden Member
Oct 16, 2019
1,815
3,634
116
Icelake taped in around June 2017 and it came out in August 2019. Without the big 10nm struggle we would haven seen Icelake early 2019. Unless there are another big 7nm delays we should get MTL in H1 2023. And Intel not long ago revealed it's a H1 2023 platform , so this makes sense. Unlike Videocardz I believe MTL will be mobile only anyways, it's a new process and from 14nm Broadwell-U and 10nm Icelake-U we do know Intel prefers mobile first. First 7nm desktop is more likely Lunar Lake.
Intel also said they'd have 7nm products for all market segments (aka mobile, desktop and server) so I don't hold the same expectations as you.

I just expect desktop to come in 2H23, around a year after Raptor Lake should.
 
  • Like
Reactions: Tlh97

coercitiv

Diamond Member
Jan 24, 2014
4,546
6,258
136
So the announcement was news for the sake of some new news? Doesn't really sound like anything to get excited about at this time.
They're just following up with their announcement from last month, they said Meteor Lake would tape in in Q2. It's not much per se, but as long as this type of announcement marks a more open Intel, it's good news in theory.

When you openly talk about your progress there's less room to cover mistakes for months and years. Let's see if they keep talking.
 

mikk

Diamond Member
May 15, 2012
3,242
1,057
136
Intel also said they'd have 7nm products for all market segments (aka mobile, desktop and server) so I don't hold the same expectations as you.

I just expect desktop to come in 2H23, around a year after Raptor Lake should.

There is another indication. Two months ago Intel accidentally made hidden documents available to the public (headline only). There only were Meteor Lake P+M documents and nothing about Meteor Lake S. Intel also said there will be 10nm for desktop two years ago but it wasn't Icelake nor Tigerlake, it will be Alder Lake. We will get 7nm for desktop but I'm not expecting MTL.
 

eek2121

Golden Member
Aug 2, 2005
1,281
1,349
136
They need to fix the detection.
They did, Intel just is not using the latest version (probably intentional).

Btw when does Alder Lake come to mobile and in what form?
The 14 core/20 thread chip tested above is supposed to be a mobile part. It is likely a late ES chip, so the results won’t represent final performance. No firm release date yet. Given that Intel is refreshing some Tiger Lake U models and has just launched Tiger Lake H, I imagine we will probably not see laptop chips until next year. It would be odd for Intel to roll out new chips only to replace them 3-5 months later.
 
  • Like
Reactions: mikk

DrMrLordX

Lifer
Apr 27, 2000
17,809
6,801
136
switch to ATX 12VO as new standard for motherboards and PSUs
That's . . . gonna be interesting. I don't see how that makes sense for DiY unless PSU OEMs do it just to reduce redundancy in their product lineup (e.g. it'll cost them more $$$ to design PSUs for the new ATX standard for OEMs and the old ATX standard for DiY)

Nevertheless, I expect with current PSU supplies being scarce, that few DiY builders are going to want to have to throw out 2-3 year-old PSUs in the 500w-1kw range (or higher!) just to adopt Alder Lake-S. Some of them might want to do it anyway, but chucking a $200+ PSU that's supposed to be good for 5-10 years for a system upgrade is gonna suuuuuuuuuuck.

and now this weird Peltier element leak. (that is hopefully just an uninspired marketing tactic)
I wouldn't make too much of that.

Wonder what the point would be given the existence of the 6+0 die.
Supply. Same reason why Intel kept pushing 14nm mobile chips when Ice Lake-U came out. They may still be having problems getting big enough dice to yield properly on 10SF and 10SFE. See reports of their being maybe 1m SoCs for the 8c TigerLake-H launch.

Intel will take full advantage of TSMC's best processes, even if they have to outbid AMD, as this is a very good way to cripple AMD long term and send them back to Bulldozer era.
Once Intel starts using TSMC as their leading-edge wafer supplier, they'll do what they're told to do. And I don't mean that in a good way.

Icelake taped in around June 2017 and it came out in August 2019. Without the big 10nm struggle we would haven seen Icelake early 2019.
Heck had Cannonlake launched on time, Ice Lake-U could have been a 2018 product. Had Broadwell launched on time, it could have been out earlier than that!
 

eek2121

Golden Member
Aug 2, 2005
1,281
1,349
136

This is literally best case for Tiger Lake-H, 1 thread bench in a 45W product. FP difference particularly impressive.
I am leaving this one alone for now.

The issue is that something is either up with Anandtech’s testing, with the “reference platform”itself, or the 11980hk chip.

From what I have seen thus far, the 11800h appears to be beating the 11980hk in both Geekbench and Cinebench, and I’ve yet to see an 11980hk hit 5ghz on Geekbench (though there aren’t many results).

I think that the fact it has a 65W TDP up mode is causing thermal issues. AT caught some of those
issues as they noted in the article, but I am wondering if the chip itself is a hot mess. I will be testing the 11800h myself when I receive my laptop, and I hope AT will receive and test more laptops soon.
 
  • Like
Reactions: mikk

eek2121

Golden Member
Aug 2, 2005
1,281
1,349
136
That's . . . gonna be interesting. I don't see how that makes sense for DiY unless PSU OEMs do it just to reduce redundancy in their product lineup (e.g. it'll cost them more $$$ to design PSUs for the new ATX standard for OEMs and the old ATX standard for DiY)

Nevertheless, I expect with current PSU supplies being scarce, that few DiY builders are going to want to have to throw out 2-3 year-old PSUs in the 500w-1kw range (or higher!) just to adopt Alder Lake-S. Some of them might want to do it anyway, but chucking a $200+ PSU that's supposed to be good for 5-10 years for a system upgrade is gonna suuuuuuuuuuck.
12VO would make PSUs cheaper, and there would be fewer, smaller connections from the PSU to the motherboard. The issue is that motherboards become more expensive to produce.

From my limited research on the subject, 12VO isn’t a bad standard. IMO it needs to happen. That, or a similar standard.

EDIT: take a look at this article: https://overclock3d.net/reviews/power_supply/atx12vo_tested_-_the_future_of_power_supplies/1

12VO simplifies everything. Fir example, since SATA power now comes from the motherboard, in theory you only need 1 cable instead of 2. That isn’t the case for the board in the article, but I can see it happening.
 
Last edited:

uzzi38

Golden Member
Oct 16, 2019
1,815
3,634
116
There is another indication. Two months ago Intel accidentally made hidden documents available to the public (headline only). There only were Meteor Lake P+M documents and nothing about Meteor Lake S. Intel also said there will be 10nm for desktop two years ago but it wasn't Icelake nor Tigerlake, it will be Alder Lake. We will get 7nm for desktop but I'm not expecting MTL.
None of those are accidental, all of those documents are password protected. We see these listings months ahead of every launch from Intel, it's far from the first time.

Intel specifically said full product stack across the board, I see no reason for them to hold back for the desktop. Just feels like the same release cadence as TGL to me. Mobile products first, higher power products half a year later.

This ain't like you Mikk, to be surprisingly conservative on Intel's roadmaps. What, are you overly worried about 7nm clocks or something?
 

DrMrLordX

Lifer
Apr 27, 2000
17,809
6,801
136
12VO would make PSUs cheaper, and there would be fewer, smaller connections from the PSU to the motherboard. The issue is that motherboards become more expensive to produce.
Well my point was . . . let's say you built a 11900k system today with a 3090 (or whatever you can actually buy, lol) and you used this bad boy:


(hard to buy these right now)

If you bought one of those in the last year or two, would you be in a hurry to get rid of it for your next system? Those things are like $230 for a refurbed one. New I think they were up to $300.

From my limited research on the subject, 12VO isn’t a bad standard. IMO it needs to happen. That, or a similar standard.
Maybe? I'm okay with the way things are from a DiY perspective. Though wire routing is a PITA with the 8-pin connectors and so forth.
 

Rigg

Senior member
May 6, 2020
218
354
96
12VO would make PSUs cheaper, and there would be fewer, smaller connections from the PSU to the motherboard. The issue is that motherboards become more expensive to produce.

From my limited research on the subject, 12VO isn’t a bad standard. IMO it needs to happen. That, or a similar standard.

EDIT: take a look at this article: https://overclock3d.net/reviews/power_supply/atx12vo_tested_-_the_future_of_power_supplies/1

12VO simplifies everything. Fir example, since SATA power now comes from the motherboard, in theory you only need 1 cable instead of 2. That isn’t the case for the board in the article, but I can see it happening.
I can see this being attractive for OEM but it sounds like a bad idea for DIY. Quality motherboards have already become more expensive in the last few generations. I'm skeptical the PSU cost savings would be passed on to consumers while increased motherboard cost almost certainly would. This could induce some serious limitations on ITX boards that are already cramped for space.
 
Last edited:

mikk

Diamond Member
May 15, 2012
3,242
1,057
136
None of those are accidental, all of those documents are password protected. We see these listings months ahead of every launch from Intel, it's far from the first time.

It was accidental and that's why we can no longer see it, they have closed this loophole and we have no longer access to this.

This ain't like you Mikk, to be surprisingly conservative on Intel's roadmaps. What, are you overly worried about 7nm clocks or something?

Volume and yield is my first concern when it comes to a new Intel process. You are really optimistic to expect they can start shipping mobile and desktop the same year. If they cannot launch Meteor Lake-S within half a year of Meteor Lake-P it's questionable if it's worth the effort, although it depends on Lunar Lake. Broadwell-U launched in January 2015 after a 6 months 14nm delay and in August 2015 they already launched Skylake-S, that's only a 7-8 months gap to the next architecture.
 

uzzi38

Golden Member
Oct 16, 2019
1,815
3,634
116
It was accidental and that's why we can no longer see it, they have closed this loophole and we have no longer access to this.
Nah, things like this are never accidental, just often overlooked. Little bit of a difference between the two.

Volume and yield is my first concern when it comes to a new Intel process. You are really optimistic to expect they can start shipping mobile and desktop the same year. If they cannot launch Meteor Lake-S within half a year of Meteor Lake-P it's questionable if it's worth the effort, although it depends on Lunar Lake. Broadwell-U launched in January 2015 after a 6 months 14nm delay and in August 2015 they already launched Skylake-S, that's only a 7-8 months gap to the next architecture.
You're absolutely right in that it depends on Lunar Lake and when that is. If Lunar Lake was originally slated to launch in late 2023 then it would indeed make no sense for MTL-S to exist. However that's a big if.

With tape-out costs for newer nodes absolutely shooting up with each progressive node shrink, and volume concerns still remaining an issue, I don't think it makes sense to target small iterative architectural improvements on a yearly basis but instead to focus on much larger architectural changes over 1.5-2 years.

But in any case, anything is possible as far as we know, lets wait and see what happens.
 

Ajay

Diamond Member
Jan 8, 2001
9,460
3,939
136
With tape-out costs for newer nodes absolutely shooting up with each progressive node shrink, and volume concerns still remaining an issue, I don't think it makes sense to target small iterative architectural improvements on a yearly basis but instead to focus on much larger architectural changes over 1.5-2 years.
Tapeout costs shooting up? This is a data transfer. Please explain.
 
  • Like
Reactions: Tlh97 and Hulk

Thala

Golden Member
Nov 12, 2014
1,266
578
136
Isn't that tape out? Tape in is sending the design to the fab, tape out is when you get the silicon back?
Tape-out and Tape-In is essentially the same - both referring to the fab getting the design data. Getting the silicon back is not associated with a "tape" at all.

ps. Sorry, I did not notice, that this was already answered at the time of my post.
 
Last edited:

IntelUser2000

Elite Member
Oct 14, 2003
7,597
2,458
136
Some thoughts on Sapphire Rapids and interconnects.

Sapphire Rapids-
If Sapphire Rapids maxes out at 56 cores, we can hope for only 60-80% performance gain, assuming everything works out. While that'll beat Milan, it'll only do it by a small amount.

56 cores = 40% more cores, Golden Cove = 20-25% more perf/clk

The thing is just putting more cores isn't enough as scaling isn't perfect so resources need to be added just to compensate for the loss. So 40% gain from 40% more cores is optimistic.

In Integer scenarios where it's representative of a uarch gain, things like HBM2 will only help in a minor way.

They really need to get the 72 core version working somehow, if not 68. If Genoa is really that big of a gain,* potentially even with 68 cores, Genoa will end up being 50%+ faster. Intel focuses more on enterprise so it might be closer there, but I can't see how they'll be in a better competitive position then they are with Icelake-SP. Maybe by a wee bit?

Interconnects-
This is more of an what-if? Based on how the 4 tiles are placed so closely on leaked Sapphire Rapids shots, perhaps they can go with "rings-of-meshes". Each tile would have it's cores connected using a mesh, but inter-tile connections would be using a ring.

*Based on how Ampere is looking, Genoa will need all the performance it can get.
 
  • Like
Reactions: lightmanek

andermans

Member
Sep 11, 2020
87
81
51
56 cores = 40% more cores, Golden Cove = 20-25% more perf/clk
Open question is how the new process and the cores change the clock at a given TDP per core level. For the top SKU they're going from 6.75 W/core for Ice Lake SP to 6.25 W/core for Sapphire Rapids, so not too much of a regression. AFAIU those power levels (~54W for 8 cores) were around the level where Ice Lake on mobile was already seeing the clocking issues from the 10nm process at the time?

Interestingly even for AMD with the 64 core -> 96 core move the TDP per core barely changes for the highest TDP models (280W for Milan and 400W cTDP for Genoa), going from 4.375 W to 4.167 W. Same for the standard TDP of 225W -> 320W, that is a 3.516 W to 3.333 W per core change. (though this overlooks all of the IO die effects wrt power).

So either going to lower TDP per core isn't that helpful for power efficiency or both AMD/Intel likely are not really trying to compete with the power efficiency of ARM servers.
 

ASK THE COMMUNITY