- Mar 3, 2017
- 1,602
- 5,788
- 136
AMD's product naming scheme has been utterly broken for years. So Ryzen 7000 could be Zen 3-based Rembrandt for desktop.
Also a WCCF's tweet...
Here is my wishful thinking of what this mystery CPU might be:
1/4 of Mi300
That would be 1 to 3 Zen 4 CCDs stacked on top of a base die. What makes me think that this is a (remote) possibility? AMD will have these Zen 4 dies that can stack on top of the base die. What is missing is a client grade base die, so this would still take some investment to bring it to market.
BorisTheBlade82
Senior member
- May 1, 2020
- 663
- 1,014
- 106
It was first brought up by Andreas Schilling as part of a Gigabyte product presentation.
My guess is an APU line that they label as 8000 series. If it really is Zen 5 launching late this year, good luck to Intel trying to catch up any time soon.
ryanjagtap
Member
- Sep 25, 2021
- 108
- 127
- 96
According to AMD's new naming scheme, the 2023 launched products will come under 7000 series. So it should be 7000G or else the whole naming scheme will become pointless.
Attachments
If it's at the end of the year I imagine it would be 8xxx. Says model year and not year year. Heh.
desrever
Member
- Nov 6, 2021
- 110
- 267
- 106
That naming scheme is only for mobile.
1. Raphael refresh (6 months before Zen5, a-la Zen2XT/3?), they probably feel having something to counter/distract from RTL refresh around the same time is best (and I would agree).
2. Desktop Phoenix, this was sort of leaked awhile back anyway, so it wouldn't exactly shock.
3. Both of the above.
4. Gigabyte misspoke.
Edit: Also pretty sure no product this year will be called Ryzen 8000 anything, mobile, desktop or anything else. And no desktop Rembrandt, but this last one wouldn't surprise me either.
Edit2: Here's random conundrum... what will AMD's 2026 lineup be called? Ryzen 10K? Or is that a reset year for them? And maybe a another "Royal Core"/2017 Zen-esque hype moment? lmao
Last edited:
BorisTheBlade82
Senior member
- May 1, 2020
- 663
- 1,014
- 106
All of that is entirely possible and plausible.
Just to add a bit of speculation - self-quote:
Maybe early q1 2024. I've said it many times but it was always clear zen 4 was a cleaner to set things in motion. I wouldn't discount gigabyte, they know a lot more than leakers do and I wouldn't discount an aic company when they mess up like this. People thought msi were trolls when they accidentally leaked and then deleted some photos of which would turn out to be the 3990x.
I have more faith in aic's knowing something than leakers or those who must have their internet taken away for being a pain in the ass like than my life is diarrhea aka mlid.
I have more faith in aic's knowing something than leakers or those who must have their internet taken away for being a pain in the ass like than my life is diarrhea aka mlid.
I thought Gigabyte did already leak everything they had xD
biostud
Lifer
- Feb 27, 2003
- 18,246
- 4,755
- 136
The Gigabyte was referring to support of future AM5 processor while launching a Ryzen Server. APU would be irrelevant to this product line.
- Mar 3, 2017
- 1,602
- 5,788
- 136
I am wondering if Execufix leak on the Zen 5 VID is still valid considering Zen 5 is now known to be not exclusively on 3nm. I speculated before that there has to be an IVR on die/package (since no TSMC process can handle this VID). AMD has good experience designing on die inductive coils since Zen 2. Challenge would be to fab it vertically instead of horizontally for more area efficiency. Should help reduce the number of Vdd pins and cut some IIR losses, some heat, improve efficiency etc. compared to DLDO.
Last edited:
- Mar 3, 2017
- 1,602
- 5,788
- 136
Yikes, looks like the rumor of Zen 5 + Zen 4D could be true after all
Reading Family ID is not sufficient anymore, now there are core level differences and the CPU Topology has to be read as well.
Reading Family ID is not sufficient anymore, now there are core level differences and the CPU Topology has to be read as well.
Last edited:
Hopefully, those cores will be in separate chiplets, not mixed. They, you would just buy a model that has only full Zen 5 cores.
- Mar 3, 2017
- 1,602
- 5,788
- 136
Same thought, I only want the Zen 5 cores, not some puny outdated Zen 4c
I would guess one 8C Zen 5 Chiplet + 16C Zen 4 Chiplet for 24C.
So going by PHX2 topology, Zen 4c would have half the L3 per Core but for 16C in one CCX the amount of L3 would still be same? That could make L3 available to a core to the SW pretty much same even though L3 per core is half. But overall around same size as Zen 4 CCD but 2x cores, if going with PHX density of 25000 MTr@178mm2
Last edited:
Same thought, I only want the Zen 5 cores, not some puny outdated Zen 4cinebenchcores.
I would guess one 8C Zen 5 Chiplet + 16C Zen 4 Chiplet for 24C.
Didn't AMD official said only 8 cores max for 1 CCD few weeks ago?
Also InstlatX64 said the Performance/Efficient core narrative is for 'Phoenix1', which benchmarks were already leaked showing only 6-8 Zen4-no-c cores.
I doubt the P/E core narrative is a big misunderstanding.
(Maybe Phoenix2 would be a Zen4+Zen4C but the leaked patent is for PHX1 only and remain unused)
Last edited:
moinmoin
Diamond Member
- Jun 1, 2017
- 4,949
- 7,658
- 136
I'd expect two 8c CCXes per Zen 4c CCD, so per CCD the L3$ amount is the same but per CCX and core it's still the half.
TRENDING THREADS
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)
- Started by DisEnchantment
- Replies: 10K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 7K
-
-
Facebook
Twitter