Discussion Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)

[Vattila]

Except for the details about the improvements in the microarchitecture, we now know pretty well what to expect with Zen 3.

The leaked presentation by AMD Senior Manager Martin Hilgeman shows that EPYC 3 "Milan" will, as promised and expected, reuse the current platform (SP3), and the system architecture and packaging looks to be the same, with the same 9-die chiplet design and the same maximum core and thread-count (no SMT-4, contrary to rumour). The biggest change revealed so far is the enlargement of the compute complex from 4 cores to 8 cores, all sharing a larger L3 cache ("32+ MB", likely to double to 64 MB, I think).

Hilgeman's slides did also show that EPYC 4 "Genoa" is in the definition phase (or was at the time of the presentation in September, at least), and will come with a new platform (SP5), with new memory support (likely DDR5).

What else do you think we will see with Zen 4? PCI-Express 5 support? Increased core-count? 4-way SMT? New packaging (interposer, 2.5D, 3D)? Integrated memory on package (HBM)?

Vote in the poll and share your thoughts!

 

[andermans]

On one hand I'm doubtful of it not being design complete at this point. On the other hand AFAIK the first early engineering samples only go to big partners some time from now, which makes a bunch of the performance leaks that have come out until now very hard to believe. (or rather anything that is more than a performance projection).

Big example here is https://chipsandcheese.com/2021/02/05/amds-past-and-future-cpus/ which claims an IPC improvement of 29% based on an engineering sample. Now I'm asking myself for every claim of 20%+ IPC improvement out there: is that a real leak or is this just someone posturing based on this (hard to believe) leak or based on their hope wrt Alderlake?

(You have some of that with every leak of course but this makes it especially hard for people that combine their own leaks with repeating other leaks)

 

[Gideon]

I don't know, and to be frank, I don't entirely care. He's said enough bollocks for me to know that he's more than happy to either make stuff up or trust things from absolutely anyone.

100% this. Usually just pure informed speculation leads to much more accurate facts that what these leakers claim.

So many of these leaks go blatantly against common industry facts that it hurts (and this goes against all of MLID, Adored and Coreteks), Things like:

• Claiming not having working silicon in the labs less than year before release
• Claiming things that would require changes to silicon (other than respins) less than a year before release.
• Claiming something has been designed but might not be released - This happens, but very rarely, as the R&D money has already been spent, it would literally have to be unsellable to get canned. (Things such as designing a 24 core Genoa and not announcing it while releasing a 16 core one makes 0 sense)
• And the big one: Knowing SKUs and pricing 6+ months pre-release (when these are the last things that get decided. Especially the pricing as it's the only thing that can be changed easily, even hours before release)

But what really grinds my gears is if they get something wrong, they almost never admit that it was (someone's) poor speculation. Near always there is the excuse of "oh it must have been canned/postponed/changed last minute".

I still remember Adored being hell-bent that Navi will release in January Q1 2019, up to late December 2018. And when it didn't happen it was just casually "postponed due to yields". It ended up "being postponed" for 7 months. I'm sure AMD had no idea of the state their yields a month before release

 

[Ajay]

Also, is there an option to hide posts without ignoring? Am getting weary about leaks lately. There is just too many and basically covers all dimensions in all the performance axes.

Me too, but not that I know of. You could write a per-processor plug-in for a browser to hide posts with certain key words, but that's too much work for me since I haven't touched java script or CSS in 15+ years.

 

[gdansk]

But why come to a thread about future processors with "speculation" in the title? How can you expect anything but rumors and leaks until it is officially announced...

 

[scineram]

From what leak do we know that Zen 4 taped out? Or is this some insider info? Certainly possible, but then I would expect AMD to announce that milestone by Q2 earnings at the latest. Maybe Computex?

 

[uzzi38]

From what leak do we know that Zen 4 taped out? Or is this some insider info? Certainly possible, but then I would expect AMD to announce that milestone by Q2 earnings at the latest. Maybe Computex?

If we get a roadmap update at Computex then it'll be then, probably. But frankly speaking, I don't think they will. What's the point, they already having crashing mindshare and enough marketing points to go off already, they don't need another one.

Because that's all that discussing the state of future designs. A marketing point.

 

[CakeMonster]

Thanks for pointing out the various BS actors who are given way much attention. I just feel less motivated to read forums and participate when I see them referenced so often.

It should be possible to discuss the technical aspects of a new generation, even quoting qualified sources speculating, but ignoring the bad actors that are just chasing clicks for a paycheck.

 

[moinmoin]

Maybe it's worth having two separate forward looking threads: One discussing confirmed data (like public roadmaps, patents etc.) and one with all the unconfirmed rumors and leaks (free for all)? I guess mods would need to enforce the rules for the former thread though.

 

[andermans]

Maybe it's worth having two separate forward looking threads: One discussing confirmed data (like public roadmaps, patents etc.) and one with all the unconfirmed rumors and leaks (free for all)? I guess mods would need to enforce the rules for the former thread though.

Be careful with patents though. They are just ideas and there is typically no indication which product they'll end up in, so from the product perspective they are just as much unconfirmed rumors.

 

[moinmoin]

Be careful with patents though. They are just ideas and there is typically no indication which product they'll end up in, so from the product perspective they are just as much unconfirmed rumors.

The difference is that discussion about those is not purely blue sky speculation but has at least some basis in reality, which is exactly what's the issue with all those youtubers. With the latter any discussion can quickly be overcrowded with pure rambling far off any efforts at educated guesses.

 

[Rigg]

I don't know, and to be frank, I don't entirely care. He's said enough bollocks for me to know that he's more than happy to either make stuff up or trust things from absolutely anyone.

You guys remember this one? I know who made it up. Hell a couple of things on their I contributed to. It was total BS designed to see who would take the bait without attempting to verify their sources, and MLID posted it in a video within just hours of receiving it.

With the exception of the memory, and the god awful reference design photoshop, everything on that fake slide was pretty much dead on to final specs and price. The specs in that slide are close enough that someone who had accurate leaked info might have been a bit perplexed by it.

MLID starts the video of by stating its most likely fake and from a random source. He does try to steel man why it could be real but never really presents it as anything more than a leak that is most likely fake.

While I'm sure there is plenty of ammo to discredit the dude, I don't think this video was all that damning.

I would hope that anyone watching youtube leakers is only doing it for the entertainment value and taking any leaks with a massive grain of salt. I enjoy reading/watching informed speculation on hardware but honestly none of it matters. People get to invested in it so they can pat themselves on the back for geting it right and dunk on the naysayers when the products finally release.

 

[Bigos]

I'm sorry, but is this a site that advertises random dudes on YT or do we try to discuss Zen 4 microarchitectures and integration here?

I for one would like not to see links to dubious sources, especially if clicking on them benefits anyone. If you want to discuss a rumor, state it here, not just a link to a place that states it. If you have a fact, state your sources.

You don't need to describe a whole YT movie in details, just enough so that we know what it is about and whether the source is reliable. I hate when I click on a source and it turns out to be a complete bullshit - I feel click-baited then. Don't expect anyone to watch through a whole movie just to realize it says about the most obvious things.

Unless you are payed by the original creator - then moral don't affect you. But doing it for free seems... useless?

 

[Vattila]

Usually just pure informed speculation leads to much more accurate facts that what these leakers claim.

Ironic, but true.

As the OP of this discussion thread, let me state that this thread was meant as a discussion thread about technology and possibilities, and not as a place to peddle rumours and unauthorised disclosures of confidential business plans.

Unfortunately, many YouTubers, Twitter members and technology news sites have specialised in leaks and try to establish themselves as reliable sources of such, which I think is dubious in many ways. It is a shame, in particular in the case of the YouTubers, since many of them are skilled content creators that show genuine interest in the history, development and future of technology and the semiconductor industry. There are many YouTube channels with superb presentations on many topics, of interest and value to enthusiasts and investors alike, who want to understand technology and the industry better. I hope we'll see more of that and less of the rumour mill.

But why come to a thread about future processors with "speculation" in the title?

To discuss and learn about CPU design trends and future possibilities.

It should be possible to discuss the technical aspects of a new generation, even quoting qualified sources speculating, but ignoring the bad actors that are just chasing clicks for a paycheck.

Exactly.

 

[Gideon]

With the exception of the memory, and the god awful reference design photoshop, everything on that fake slide was pretty much dead on to final specs and price. The specs in that slide are close enough that someone who had accurate leaked info might have been a bit perplexed by it.

The Combo of 14GB, GDDR6 and 1008 GB/s means a (stupidly wide) 448bit memory bus with 18 Gbps GDDR6 (that hasn't really become a thing). It would not pass my BS meter unless there were multiple sources claiming that (which there clearly weren't since it was made up)

 

[DisEnchantment]

AMD HSA is here

[RFC PATCH 0/5] Support DEVICE_GENERIC memory in migrate_vma_*

lists.freedesktop.org

AMD is building a system architecture for the Frontier supercomputer with
a coherent interconnect between CPUs and GPUs. This hardware architecture
allows the CPUs to coherently access GPU device memory. We have hardware
in our labs and we are working with our partner HPE on the BIOS, firmware
and software for delivery to the DOE.

The system BIOS advertises the GPU device memory (aka VRAM) as SPM
(special purpose memory) in the UEFI system address map. The amdgpu driver
looks it up with lookup_resource and registers it with devmap as
MEMORY_DEVICE_GENERIC using devm_memremap_pages.

Now we're trying to migrate data to and from that memory using the
migrate_vma_* helpers so we can support page-based migration in our
unified memory allocations, while also supporting CPU access to those
pages.

 

[DrMrLordX]

Earlier than expected if true, the last rumors pointed towards late 2022. It's a bit early to predict anyways, a lot can happen in the next months.

Assuming AMD doesn't break with their admittedly-turgid pace of 15-18 months between major releases, latest date for Raphael would be somewhere in Q2 2022. Earliest would be Q1 2022.

 

[rainy]

Assuming AMD doesn't break with their admittedly-turgid pace of 15-18 months between major releases, latest date for Raphael would be somewhere in Q2 2022. Earliest would be Q1 2022.

I see a perfect date for release of Zen 4 - 5nm on 5th May next year, similarly to Zen 2 (7nm on 7th July 2019).

Btw, it would be exactly 18 months since Zen 3 debut.

 

[Ajay]

Assuming AMD doesn't break with their admittedly-turgid pace of 15-18 months between major releases, latest date for Raphael would be somewhere in Q2 2022. Earliest would be Q1 2022.

I really wonder if COVID pushed out the Zen4 product release dates. That, or we are being gas lit on the rumored late release dates. AMD only puts out vague roadmaps publicly - so we'll have to wait to a really solid leak comes from an OEM :sad_panda:

 

[Mopetar]

There are a lot of other factors that might go into a product release date and plenty of other problems that can lead to delays so company's are always going to want to have some flexibility. Unless there are upstream delays (e.g. TSMC falling a bit behind with rolling out a new node that some product will use) most of the pandemic related supply issues hit existing production the same as they would future products. The ABF substrate shortage applies to current chips just the same as the next generation chips, so if it's something like that then it seems to make more sense to transition to the new chip if there's going to be shortages either way.

 

[DrMrLordX]

I really wonder if COVID pushed out the Zen4 product release dates.

Really depends on what measures they took (if any) to keep teams working.

 

[eek2121]

Assuming AMD doesn't break with their admittedly-turgid pace of 15-18 months between major releases, latest date for Raphael would be somewhere in Q2 2022. Earliest would be Q1 2022.

I see a perfect date for release of Zen 4 - 5nm on 5th May next year, similarly to Zen 2 (7nm on 7th July 2019).

Btw, it would be exactly 18 months since Zen 3 debut.

Except Warhol is still somehow rumored to be the next release.

PS if you need a 5950X and live in the US, find the nearest microcenter, check the stock levels on the website, and do a road trip. The one I visited had dozens in stock. The same goes for all the other chips.

I will be installing mine on monday.

 

[IntelUser2000]

I really wonder if COVID pushed out the Zen4 product release dates. That, or we are being gas lit on the rumored late release dates. AMD only puts out vague roadmaps publicly - so we'll have to wait to a really solid leak comes from an OEM :sad_panda:

I doubt it added any significant amount of time.

Remember modern chips are super complicated things to manufacture and validate. The teams are extremely talented to create such complex devices without having bugs left and right.

Things like chiplets, on package memory, EMIB will further complicate the process.

 

[DrMrLordX]

Except Warhol is still somehow rumored to be the next release.

It is unknown what, if any, place Warhol has in AMD's release schedule. If it's a full release on the level of Matisse or Vermeer, then it would put Raphael in 2023. Which is ridiculous.

 

[Asterox]

Hm, this is no doubt a bold IHS design.

https://twitter.com/x/status/1399076515837530112

 

[LP-ZX100C]

Hm, this is no doubt a bold IHS design.

https://twitter.com/x/status/1399076515837530112

It's a bit too thick the IHS , for my liking .