• 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."

Speculation: Ryzen 4000 series/Zen 3

Page 173 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

A///

Senior member
Feb 24, 2017
771
516
106
I'm assuming Warhol, if real, would introduce an entirely new chipset; regardless, it isn't stupid if AMD thinks they'll have no credible competition. Gotta milk that cow for all its worth. Moo.
Mozart or whatever his name on Twitter is seems to think that RKL will be out with a 5.5 SCC, and that Alderlake will follow with some refined designs.
 

A///

Senior member
Feb 24, 2017
771
516
106
It would be dumb, but they also did the XT refresh which was an absolute joke.

Realistically, an N6 Zen 3 shrink and bugfix a la Zen+ would be the best option.
That's assuming there is a bug to fix. N6 makes some sense, I suppose. Technically, they could offer some of the non-controversial tech on it. USB4, for example.
 
Feb 17, 2020
38
67
51
That's assuming there is a bug to fix. N6 makes some sense, I suppose. Technically, they could offer some of the non-controversial tech on it. USB4, for example.
There are always bugs to fix. Ever look at an errata sheet? That's just the hardware bugs you decide to live with.
 

A///

Senior member
Feb 24, 2017
771
516
106
@DrMrLordX I wonder if AMD will try anything funky like PGA to PGA adapters from the old Socket 5 and 7 eras.

There are always bugs to fix. Ever look at an errata sheet? That's just the hardware bugs you decide to live with.
Yes, but that's beside the point. Most bugs are a non-issue. Major bugs are.
 
Last edited:

lightmanek

Member
Feb 19, 2017
169
220
86
The rumor about Infinity Cache makes me wonder if AMD is going to use it as a backbone of creating chiplet GPU-GPU interconnect.
Complexity is enormous and even 128MB is not enough to share frame data, but if used in clever way ...

The name Infinity Cache, if looked as an extension of Infinity Fabric which job is to connect multiple CCX's and CCD's to create one CPU, can be interpreted as cache logic to glue multiple GPU dies.

I know it makes little sense as even AMD stated that chiplet GPUs are only feasible for compute/Ai workloads and architectures like CDNA, but what if they were Jebaiting us 😄

Too early, need my coffee, I will shut up!
 

DrMrLordX

Lifer
Apr 27, 2000
16,371
5,282
136
I think a zen3 ddr5 release would be all about getting the platfrom in place and ready for the real launch. Hopefully it would just be mid cycle with zen4 with the same i/o die/chipset coming out at the standard year + a bit interval from initial zen3 release.
One thing is clear: with Vermeer being so late in 2020, it's almost inevitable that Zen4 desktop (Raphael?) will be a 2022 chip. AMD missed on their roadmap. An interim chip might whet people's appetites a bit and help distract from the fact that they aren't staying on-track.

Mozart or whatever his name on Twitter is seems to think that RKL will be out with a 5.5 SCC, and that Alderlake will follow with some refined designs.
I doubt it would help much, especially if that's a TVB clockspeed.

@DrMrLordX I wonder if AMD will try anything funky like PGA to PGA adapters from the old Socket 5 and 7 eras.
I hope not. That would be awkward.
 
  • Like
Reactions: A///

A///

Senior member
Feb 24, 2017
771
516
106
Definitely TVB, @DrMrLordX. No way they can get more than a core or two at those speeds. Hell, I'd love to see wattage and thermals on that. As for the adapter, it's plausible, but given it's hardly done for laptops, and that's for scrapped out ones not from OEMs, it's very unlikely.

I'm sure if Warhol is real, there is a logic method to introduce it as a stepping stone.
 

moinmoin

Golden Member
Jun 1, 2017
1,781
1,789
106
No, it's a joke. You're talking maybe 100 MHz of uplift, which is basically negligible.
It's a typical "the competition makes minor upgrades and the market still eats it up, so let's do that as well" move. It's only a joke in context of AMD's own products within the last 3 years, which is quite telling actually.

AMD missed on their roadmap.
12 vs. 15 months cadence. If Zen 3 does launch in October they missed 15 months by just one month which could be attributed to the pandemic.
 

turtile

Senior member
Aug 19, 2014
482
91
91
That's assuming there is a bug to fix. N6 makes some sense, I suppose. Technically, they could offer some of the non-controversial tech on it. USB4, for example.
N6 doesn't make anysense. Unless the new 7nm they are using is EUV, they are not design compatible. Given Zen 4 will be out in a year or so, why bother? AMD is also using a customized 5nm from TSMC.

I don't recall AMD ever stating they were on a strict 12 month cadence.
Yes, I think they were targeting 12-15 months.
 
  • Like
Reactions: Tlh97

Ajay

Diamond Member
Jan 8, 2001
7,260
2,491
136
N6 doesn't make anysense. Unless the new 7nm they are using is EUV, they are not design compatible. Given Zen 4 will be out in a year or so, why bother? AMD is also using a customized 5nm from TSMC.



Yes, I think they were targeting 12-15 months.
I could be wrong, but I don't recall AMD saying anything other than 'by the end of 2020'. Don't know what their internal timeline was - have they revealed that?
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,188
691
136
N6 doesn't make anysense. Unless the new 7nm they are using is EUV, they are not design compatible.
There is no indication of N7+ at AMD at all. However, there were plenty of N6 references...
"In a team take charge of products of AMD in N7/N6 node"

¯\_(ツ)_/¯

Vermeer except for a couple blocks is basically the same die as Matisse.
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,188
691
136
Man, stop doing this. Zen3 didn't have all this development time put into it to change 'a couple of blocks'.
First Zen3 was canned sometime before 2H18. (Team A pushed first to Zen5)
The final Zen3 was then split into two designs: On a 7nm family node which is a fake Zen3(cut short design via olde team) and on a 5nm family node which is the real Zen3(longer design phase via another team). Zen(A) -> Zen2(B) -> Fake Zen3(B) -> Real Zen3(C)-N5 -> Zen4(C)-N5+ -> Zen5(A+B)-N3

That is the general gist that I have been given.

If AMD didn't have multiple concurrent design teams, Starship(SSP packages) would only be packin' 48-cores max.

Don't get me on Team D & E which is going to appear for 6 & 7; this early on.
 
Last edited:

DrMrLordX

Lifer
Apr 27, 2000
16,371
5,282
136
12 vs. 15 months cadence. If Zen 3 does launch in October they missed 15 months by just one month which could be attributed to the pandemic.
I don't recall AMD ever stating they were on a strict 12 month cadence.
Ahem


Zen4 in 2022 is a miss, plain and simple. It's time to stop making excuses for AMD while bashing Intel for their roadmap mishaps. They set a goal, and they missed it.
 
  • Like
Reactions: Lodix and ozzy702

CakeMonster

Senior member
Nov 22, 2012
953
60
91
And Intel has max 8 core on their upcoming CPU design? I think I read that somewhere here. That means that Zen3 will face off against the 10xxx series for the next 2 years? Sounds boring from both camps with regards to innovation.... :/
 

DrMrLordX

Lifer
Apr 27, 2000
16,371
5,282
136
And Intel has max 8 core on their upcoming CPU design? I think I read that somewhere here. That means that Zen3 will face off against the 10xxx series for the next 2 years? Sounds boring from both camps with regards to innovation.... :/
Rocket Lake-S will indeed be limited to 8 cores. Intel is still using theri 14nm process and the cores of Rocket Lake-S should be a good bit larger than the Skylake-derived cores of Comet Lake, so in order to prevent the die from becoming enormous, Intel has to cut down on the # of cores. It may also prevent runaway thermals. We don't know (yet).

Allegedly Intel has Alder Lake-S coming next year, so it will actually be the product that Zen3 will have to face later in its lifespan, assuming Alder Lake-S launches on time (it probably won't, though).
 

Hitman928

Platinum Member
Apr 15, 2012
2,778
2,126
136
Ahem


Zen4 in 2022 is a miss, plain and simple. It's time to stop making excuses for AMD while bashing Intel for their roadmap mishaps. They set a goal, and they missed it.
I believe you are misinterpreting those slides. According to your interpretation, in March 2020, AMD was promising they'd go back in time and release the first Zen 3 product in 2019.



Edit: In other words, in that presentation AMD was promising Zen 4 would roll out some time in 2022 which they should have no problem achieving.

 

DrMrLordX

Lifer
Apr 27, 2000
16,371
5,282
136
I believe you are misinterpreting those slides. According to your interpretation, in March 2020, AMD was promising they'd go back in time and release the first Zen 3 product in 2019.
Milan was sampling last year, if I recall correctly. I think you're misinterpreting the first slide.

edit: now that I think of it, where did you even get that first slide? It isn't from the slide deck I linked.
 
Last edited:

Karnak

Member
Jan 5, 2017
190
213
86
Zen4 in 2022 is a miss, plain and simple. It's time to stop making excuses for AMD while bashing Intel for their roadmap mishaps. They set a goal, and they missed it.
Nope. Quoting Mark Papermaster:

Dr. Ian Cutress: So far AMD’s rate of new products is on track to produce a new core almost every year. The roadmaps quite proudly showcase Zen 3 as almost ready, Zen 4 in development, and Zen 5 further out. Is this cadence sustainable?
Mark Papermaster: We’re on a 12-18 month cadence, and we believe that is sustainable. It’s what the industry demands from us.
 

DrMrLordX

Lifer
Apr 27, 2000
16,371
5,282
136
@Karnak

Papermaster was fudging. What he said basically contradicted their own roadmap. I'm not the only one that was expecting Zen4 by the end of 2021:


Technically, if AMD can get Genoa into the hands of a limited set of partners by the end of 2021, maybe the roadmap is still on-track. But the client map from that slide release shows Zen3 up through the end of 2020, so there's no reason why they wouldn't have Zen4 client in 2021 . . .
 
Last edited:

turtile

Senior member
Aug 19, 2014
482
91
91
@Karnak

Papermaster was fudging. What he said basically contradicted their own roadmap. I'm not the only one that was expecting Zen4 by the end of 2021:


Technically, if AMD can get Genoa into the hands of a limited set of partners by the end of 2021, maybe the roadmap is still on-track. But the client map from that slide release shows Zen3 up through the end of 2020, so there's no reason why they wouldn't have Zen4 client in 2021 . . .
He didn't. The roadmap shows Zen 4 scheduled for release in 2022. It also shows no new CPU releases for EPYC in 2021. The desktop roadmap just posted above shows only Zen 3 through 2021. 2021 will just be APU/GPU releases.
 
  • Like
Reactions: Tlh97

MrTeal

Platinum Member
Dec 7, 2003
2,837
404
126
Ahem


Zen4 in 2022 is a miss, plain and simple. It's time to stop making excuses for AMD while bashing Intel for their roadmap mishaps. They set a goal, and they missed it.
The article you linked has the following paragraph in it, immediately after the server roadmap.

Perhaps most impressive, AMD's roadmaps indicate the company will have its EPYC Genoa processors fully on the market by the end of 2022. Those chips will come with the 5nm process and Zen 4 architecture, and if AMD can continue to execute on time, that represents an impressively huge jump forward in a relatively short span. We can also expect that AMD's Zen 4 processors for the consumer market will come with the 5nm process as well.
 

ASK THE COMMUNITY