Discussion AMD Cezanne/Zen 3 APU Speculation and Discussion

[dr1337]

https://twitter.com/i/web/status/1292844896374718470

Fresh leak out today, not much is known but at least 8cu's is confirmed. Probably an engineering sample, core count is unknown and clocks may not be final.

This is very interesting to me because cezanne is seemingly 8cu only, and it seems unlikely to me that AMD could squeeze any more performance out of vega. A cpu only upgrade of renoir may be lackluster compared to tigerlake's quite large GPU.

What do you guys think? Will zen 3 be a large enough improvement in APU form? Will it have full cache? Are there more than 8cus? Has AMD truly evolved vega yet again or is it more like rdna?

 

[NTMBK]

Remember, there is a non-trivial amount of interest in uSFF PCs that can be used in entertainment centers for gaming on living room TVs. Getting something in there that can competently play things at 1080p is a big deal. We're very close with the 5700g, with things like FSR upsampling 900p to 1080p with reasonable detail levels being playable in most situations. It's not unreasonable for the next gen of RDNA2 based iGPU equipped APUs to be able to hit decent 1080p/60 with decent quality settings. That makes for a lot of very functional laptops and STBs without needing a dGPU.

Okay, what are the sales numbers on that? Because it feels like a tiny niche.

 

[Mopetar]

That really depends on what AMD would charge for a part like that, because you could probably make a decent Steam box around something like that. Given the impossibility of getting a console right now, let alone at no markup, I think there are plenty of people who would drop a few hundred dollars on something like that.

After all, wouldn't it be the same market as the Xbox series S is trying to carve out?

 

[moinmoin]

That really depends on what AMD would charge for a part like that, because you could probably make a decent Steam box around something like that.

If Steam Deck sees any prolonged popularity beyond actual launch I expect OEMs to jump on the SFF bandwagon.

 

[DrMrLordX]

If Steam Deck sees any prolonged popularity beyond actual launch I expect OEMs to jump on the SFF bandwagon.

Steam Deck is portable. SFFs are not.

 

[moinmoin]

Steam Deck is portable. SFFs are not.

That's not the point.

 

[DrMrLordX]

That's not the point.

I mean . . . it really is, isn't it? Why would you buy a Steam Deck if you couldn't take it with you?

 

[moinmoin]

I mean . . . it really is, isn't it? Why would you buy a Steam Deck if you couldn't take it with you?

No OEM would try to compete against Steam Deck, especially not at the prices Valve is calling. What are you on about?

 

[DrMrLordX]

No OEM would try to compete against Steam Deck, especially not at the prices Valve is calling. What are you on about?

I mean the Steam Deck won't have anything to do with OEM interest in SFFs whatsoever.

 

[moinmoin]

I mean the Steam Deck won't have anything to do with OEM interest in SFFs whatsoever.

🤦

I guess I have to spell it out for you. Steam Deck is the result of the lessons Valve has taken from Steam Machines (remember those?). The goal with Steam Machines was a simplified console-like PC experience. For PC OEMs Steam Machines would be an additional market they could sell PCs in. The failure of Steam Machines essentially was due to 1) lack of compatibility (unless you go Windows which defeats the simplification), 2) lack of performance and 3) lack of a leading product prominently selling that platform.

Point 1 is handled by Proton, with Valve promising the remaining major blocker being anti-cheat protection getting cared of before the Deck's launch. Point 2 is slowly but surely covered by APUs (this thread's topic). Point 3 is this round Steam Deck is taking the role of the leading product prominently selling the Steam Big Picture mode/Steam OS platform.

And how could OEMs profit of a possible popularity of the Steam Big Picture mode/Steam OS platform without directly competing against Steam Deck? Selling hardware running the platform in other form factors, like SFF...

 

[Mopetar]

I mean . . . it really is, isn't it? Why would you buy a Steam Deck if you couldn't take it with you?

Tech lust.

I have a Switch, but most of the time it's just plugged in to the TV. I'd probably do the same with a Steam Deck to be honest.

 

[DrMrLordX]

🤦

I guess I have to spell it out for you. Steam Deck is the result of the lessons Valve has taken from Steam Machines (remember those?). The goal with Steam Machines was a simplified console-like PC experience. For PC OEMs Steam Machines would be an additional market they could sell PCs in. The failure of Steam Machines essentially was due to 1) lack of compatibility (unless you go Windows which defeats the simplification), 2) lack of performance and 3) lack of a leading product prominently selling that platform.

Point 1 is handled by Proton, with Valve promising the remaining major blocker being anti-cheat protection getting cared of before the Deck's launch. Point 2 is slowly but surely covered by APUs (this thread's topic). Point 3 is this round Steam Deck is taking the role of the leading product prominently selling the Steam Big Picture mode/Steam OS platform.

And how could OEMs profit of a possible popularity of the Steam Big Picture mode/Steam OS platform without directly competing against Steam Deck? Selling hardware running the platform in other form factors, like SFF...

I fundamentally disagree with all of that, which was my original point. People aren't buying Steam Big Picture Mode/Steam OS. They're buying a portable game machine.

Put Steam Deck in SFF format and people aren't gonna line up to buy it. Proton and APUs aren't changing that picture.

Tech lust.

I have a Switch, but most of the time it's just plugged in to the TV. I'd probably do the same with a Steam Deck to be honest.

SFFs options exist now. They're still niche.

 

[Shivansps]

The Aya Neo existed before the Steam Deck whiout any big company behind, petty sure others can do the same, the question would be if the market is big enoght for such a device, i guess well see. RMB is going to outperform the Steam Deck for sure, so there is motive to do clones.

As for SFF, SFF or ITX mb with soldered CPUs is were OEMs dump what they were unable to sell with notebooks. It is going to take a while for RMB to show up, they would have to get rid of any remaining Renoir/Cezanne first.

 

[LightningZ71]

I argue that the SFF space remains RELATIVELY niche precisely because it has lacked capability instead of a lack of any sort of demand. We are on the cusp of REASONABLE capability (1080p at decent frame rates) from APU based SFF boxes. I still firmly believe that, with that particular market segment, if you build it, they will come.

 

[moinmoin]

People aren't buying Steam Big Picture Mode/Steam OS. They're buying a portable game machine.

Just like people aren't buying Windows, they are buying PCs. Gotcha.

 

[JasonLD]

Just like people aren't buying Windows, they are buying PCs. Gotcha.

I agree with DrMrLordX here. Biggest selling point of Steam Deck is being the portable PC gaming device. I don't see the Steam Deck kickstarting the popularity of SFF PCs, but I could see more OEMs trying to release Steam Deck like devices.

 

[Insert_Nickname]

We are on the cusp of REASONABLE capability (1080p at decent frame rates) from APU based SFF boxes.

ASUS already make some pretty great SFF APU boxes:

https://www.asus.com/Displays-Desktops/Mini-PCs/PN-PB-series/Mini-PC-PN51/

Replace with a Rembrandt version, and things get even better. It really is amazing what can be achieved in a NUC-like formfactor.

If people ask for a desktop, these (and their Intel counterparts) will be my first choice going forward. Unless there are special considerations.

 

[moinmoin]

I agree with DrMrLordX here. Biggest selling point of Steam Deck is being the portable PC gaming device. I don't see the Steam Deck kickstarting the popularity of SFF PCs, but I could see more OEMs trying to release Steam Deck like devices.

I can also see OEMs being as stupid as that, releasing more Steam Deck like portable PCs at the price of GPD Win units. That just won't be a potential mass market, instead serving (and potentially increasing) an existing premium niche market.

I just think people are losing sight of Valve's end goal there. It's not making any of their own hardware ventures mass market successes, for such there are better ways to go about that than selling them online on one store. All the hardware are essentially examples of open implementations intended to boost the open PC ecosystem, of which Valve hopes other PC hardware manufacturers pick them up. From that happening Valve directly profits by having the dominant gaming store front for PC. So if they manage to push some hardware that gains mainstream popularity, is an open PC and as such runs Steam and its games, Valve will profit hugely. And the intersection between PC hardware with sufficient performance at a mass market compatible pricing is going to be a SFF PC, imo no way around that.

 

[JasonLD]

And the intersection between PC hardware with sufficient performance at a mass market compatible pricing is going to be a SFF PC, imo no way around that.

Consoles are just so much better value compared to SFF PCs around mass market price.

 

[blckgrffn]

I've been resisting commenting on this but here goes....

Huge media PC guy here. Built many of them, huge proponent of them. Did tons of different combinations of MCE boxes on XP and 7. Always had nice, quiet towers next to my TV's in all rooms from mid 2000's until last year.

But it's dead. I've decommissioned all of my HTPCs. All of my families HTPCs. All of my friends HTPCs. Without MCE, the "TV attached to my TV all the time" is just expensive and cumbersome. This is before we try to get Dolby Vision, high quality audio codecs all the rest of that garbage to work.

It's dead to me. I have an overly capable Plex PC and the right clients for my media at each TV. The only way I can justify that PC now is that its a simulation PC for my indoor biking too.

There are too many great options. Consoles - especially the new ones are fast and supremely capable. Huge QoL improvement over the last gen. Embarrassingly powerful next to any APU likely released in the foreseeable future.

Capable smart TVs. Rokus. Apple TVs. Etc. They have good remotes, remote listening, tons of apps, slick interfaces, and are (relatively) super cheap.

Want a more PC experience? Ever improving SBC's make it tough to justify anything more than $100 as solution.

Finally, building a bespoke PC to hook up to your TV? It's almost always more expensive to go SFF. Let alone the folks I know who just hook an old laptop up to their TV when they want that out of convenience. The longevity gains of the last decade really make that easy to do for some impromptu Steam Big Screen gaming. Plus you can use a laptop as, you know, a laptop.

It's a niche of a niche of niche.

That's all I can say about it, really. There will always be those that are drawn to SFF for their elegance, I get that. But I think most people want bang for their buck and nearly everything is better, especially eyed with any look to the longer term.

 

[moinmoin]

Consoles are just so much better value compared to SFF PCs around mass market price.

At launch they most certainly are. But they also are never going to be anywhere close to the openness PC systems offer. And that openness is not only an advantage for customers, but also manufacturers and developers.

Did tons of different combinations of MCE boxes on XP and 7.

I feel you. Windows never has been suitable for this use case.

 

[JasonLD]

At launch they most certainly are. But they also are never going to be anywhere close to the openness PC systems offer. And that openness is not only an advantage for customers, but also manufacturers and developers.

I mean, you can argue all day about how open the PC platform compared to consoles, but isn't really a strong selling point against consoles. Majority of consumers doesn't give a damn about openness of the platform over convenience and ease to use factor.

 

[Mopetar]

If you just want something to get the job done and don't have any niche or esoteric needs a console or some other device is likely good enough. On the other hand if this is an excuse to tinker and you don't care about the extra time investment then the PC is going to be better.

It really just comes down to how much of a hobby it is for most people because even if they have simple needs that a commercial product already solves, some people like to toy around with computers and the satisfaction of building their own box is greater than any of the content it will play.

 

[NTMBK]

I argue that the SFF space remains RELATIVELY niche precisely because it has lacked capability instead of a lack of any sort of demand. We are on the cusp of REASONABLE capability (1080p at decent frame rates) from APU based SFF boxes. I still firmly believe that, with that particular market segment, if you build it, they will come.

You can get a massively better experience by not limiting yourself to an APU. Slap in a GPU, and performance >doubles. Most people care more about good looking games, rather than a slightly smaller box.

 

[NTMBK]

Anyway, back to Zen 3... some stuff from Hotchips I hadn't seen before. AMD are using a bidirectional ringbus inside the CCX for communication:

 

[Insert_Nickname]

On the other hand if this is an excuse to tinker

Wait. Does one need an excuse for that?

and you don't care about the extra time investment then the PC is going to be better.

For me, tinkering with PCs is a hobby. Among others. So there isn't a finite time limit, other then spare time available.

I'm done doing it for other people though. Most needs are perfectly served with an OEM system today. Which again means less time doing support on stuff, so leaves even more time for tinkering with my own systems. Win-win as I see it.