PlayStation 4 Pro SoC Discussion

[Sweepr]

So what do we know about the rumoured PlayStation 4.5? In the wake of last week's post-GDC outing of the hardware by Kotaku, we have independently established that it's real and that Sony's R&D labs have prototype devices, and we also have more than one source referring to it as PlayStation 4K, the name we'll be using for now. And this is where things become slightly strange - because while more GPU power is being offered to developers, realistically it is nowhere near enough to provide native 4K gaming at the same quality level as current 1080p titles. The full extent of the spec is a current focus of enquiry for us, but realistically, it is simply impossible to cram the equivalent of today's top-end PC graphics hardware into a console-sized, mass-market box.

Everything we've heard positions PlayStation 4K as a machine capable of playing current and next-generation ultra HD media, while also offering support for other aspects of the 4K spec, such as high-dynamic range and a wider colour gamut - aspects of the 4K spec that could be introduced to gaming. However, in terms of additional computational power, we've got be realistic about what Sony can deliver with a mid-generation refresh.

Option #1: A new, more powerful PlayStation

What we might expect:

An APU with a higher-end Polaris would push graphics on - a 2x performance boost in GPU power compared to PS4 is achievable in a console form-factor. Possibly more - we really need to see the desktop PC equivalent parts first.
By default, Polaris has support for 4K, HDMI 2.0, HDCP 2.2 and HDR.
Console would not be cheap owing to the size of the processor - conceivably on par with PS4's £359/$399 launch price.
Possibly higher depending on how much Sony pushes the boat out in terms of processor size and memory allocation.
This PS4K could co-exist with a cheaper 'PS4 Slim' based on the older APU, again using 14nm/16nm technology.

The bottom line: In 2017, production technologies on 14/16nm should be more mature, and a significantly more powerful PlayStation could be released. But the pace of technological progress in the PC space isn't as rapid as it is on mobile. It took four dramatic leaps in chip production technology to make the generation leap between PS3 and PS4 possible. By 2017, there will have only been one viable jump in fabrication technology available to console manufacturers, and expectations should be limited. Actual 4K games will be in very short supply.

Option #2: PlayStation 4 evolved

What we might expect:

We'd see a new PlayStation playing host to the same titles as the current one, but with visual improvements.
Conceivably, older games may run more smoothly by default, or could be patched to access the newer hardware.
Resolution could be pushed beyond 1080p and could look good on a 4K screen, but native UHD visuals for triple-A titles are off the table.
Wouldn't be as expensive as the first option.

State-of-the-art 4K media support, but harder to sell to all but the most hardcore gamers.
The bottom line: Scaling up and enhancing the current PlayStation doesn't sound particularly exciting and we do have to wonder what the point would be of a PS4K with very little chance of servicing 4K gaming. However, this is the design that stands the best chance of offering a decent bump to system capabilities without introducing too much in the way of compatibility issues. Our gut feeling right now (nothing more) is that this is the form the current prototypes take.

Option #3: PlayStation plus

What we might expect:

Full compatibility with 4K screens, including next-gen media.
Complete 'no worries' compatibility with existing PlayStation 4 library.
HDR support for gaming on 4K displays - even if gaming resolution remains at 1080p.
Opportunity to lower cost as 14/16nm chips become cheaper.
Unlikely to alienate the existing userbase - any performance upgrade would be more 'nice to have' as opposed to essential.
The bottom line: PS4K could simply be a smaller, leaner version of the existing hardware, with revised 4K media and (potentially) HDR support. The option exists to overclock the existing architecture and perhaps unlock the full potential of PS4's APU. It would be like an 'Elite' version of the existing console - nice to have, but not a generational or even half-generational upgrade in terms of processing power. The question is, assuming a 2017 launch, would that be enough?

www.eurogamer.net/articles/digitalf...ystation-4k-is-real-but-why-is-sony-making-it


With M$ also rumoured to release a faster Xbox, looks like mid-generation refreshes are here to stay. I choose Option #3 - (slightly, if at all) upgraded hardware without breaking PS4 game library compatibility (higher clocks) and improved 4K/VR multimedia support.


Edit (19th April):

Revealed: Sony's plan for PlayStation 4K - codename Neo
The spec leak is genuine - here's everything you need to know.

We have been sitting on a number of details awaiting a second source before going to press, but events have overtaken us somewhat - Sony is now openly sharing this specification with developers and while Giant Bomb beat us to the punch, we have access to the same documentation. There is no doubt - this is real. This is the new, more powerful PlayStation 4.

The release window is unclear, but the schedule for hardware roll-out to developers is black and white: development kits prototype are on their way to studios now. A test kit (debug station, if you like) housed within a non-final chassis - which Sony is asking developers not to show - follows shortly. A second-gen test kit, again not based on the actual retail shell, goes out in June. Sony gives more intensive Neo briefings at its DevCon event in in May, while code submission for Neo-compatible titles begins in August.

Our concern was that PS4's low-level APIs may not be compatible with the newer architecture, meaning problems running older games, but it seems that this is not an issue. And the good news here is that Polaris' efficiency improvements could add still further to the expected increase in performance. Certainly, we should expect to see cumulative improvements to memory compression, which should help us to get more out of the constricted 256-bit GDDR5 interface. To the best of our knowledge, there were no such technologies in place on the original PlayStation 4.

Well, according to Sony's own documents, there is a focus on delivering 4K gaming content, though upscaling to UHD resolution is likely. Owners of 1080p screens can expect benefits too, explicitly stated as:

Higher frame-rates
More stable frame-rates
Improved graphics fidelity
Additional graphics features
Sony describes 'forward compatibility' via patches, allowing developers to revisit their existing PS4 library and add Neo features to existing games. Sony has opened up more memory for Neo titles too. Quite why this extra RAM can't be given to games running in Base mode isn't revealed but Sony states that Neo titles will have access to 5.5GB of memory, with 512MB "only available" for Neo mode. Sony also reveals that the background media functions of the PS4 "might be" expanded - such as the addition of 1080p gameplay recording.

www.eurogamer.net/articles/digitalfoundry-2016-sonys-plan-for-playstation-4k-neo-revealed

 

[ShintaiDK]

Benchlife mentioned something about 256MB and a decoder chip. Could simply be bluray ready PS4 with the external VR part integrated as well.

I cant imagine they are going to fracture the PS4 segment gaming wise. Because that's the whole point of a console.

 

[SPBHM]

yes I guess you could be something as boring (and that makes the most sense) as a PS4 with better support for 4K displays (but rendering at the same res as the current one) for blu-rays and also VR without the external box

 

[raghu78]

I think the current gen consoles are quite limited by the CPU portion. If Sony is really serious about improving the gaming experience they should go for a true next gen console with Zen/Polaris/HBM2. I actually think the 14nm node is not the right candidate. I would say the TSMC 7nm node or Samsung/GF equivalent would provide them enough transistors to make a true 4K/VR console. I think Sony and Microsoft are better off waiting till late 2019 or 2020 to release a true next gen console. The process nodes available at that time would provide 4 times the transistors for roughly same or lesser power. That would allow a huge improvement in performance and make 4K/VR a great experience.

 

[Silverforce11]

As said in the other threads.

Zen + Polaris + 8GB HBM2 on an interposer. A total SOC, no need system ram components. All they need is a HDD and the unit.

Fury X class performance currently delivers ~30 fps in Ultra in modern games. Dial it down and it's suddenly very capable, ~45 fps.

Polaris 10 if it's Fury X+ will easily deliver 60 fps 4K Medium/High settings with optimizations and HSA for improved performance due to on-interposer links between all components.

The biggest issue is backwards compatibility. But it's still x86 and GCN.

If either company do it, the other must as well. Given Eurogamer claims they know prototypes exist, then we're already well along the process.

2017 with 14FF, very do-able.

 

[Yuriman]

^ Such a chip would be a bit expensive for consoles, where this generation both have gone with relatively "low end" hardware (Cat cores + HD7770 or ~7850), presumably so as to not lose money on each console sold, and keep the selling price low. Upgrading to big cores + the equivalent of a $500 GPU does not seem reasonable in this context, especially on a more expensive process.

 

[Sweepr]

If there is a GPU upgrade (big if) - they could use Polaris 10 for twice as many EUs as the PS4 while keeping 256-bit GDDR5, coupled with >2GHz Puma cores.

Benchlife mentioned something about 256MB and a decoder chip. Could simply be bluray ready PS4 with the external VR part integrated as well.

I cant imagine they are going to fracture the PS4 segment gaming wise. Because that's the whole point of a console.

Sounds more likely.

 

[frozentundra123456]

Actually there is a large thread about this already in the VC & G section. As usual, one small blurb gets the hype train full steam ahead. I think it will be several years before we see anything like true 4K *gaming* performance in the consoles. I mean, we dont even have Zen yet, and Zen APUs are even further in the future, with our without HBM, and we have no idea about performance or cost.

 

[Sweepr]

Actually there is a large thread about this already in the VC & G section. As usual, one small blurb gets the hype train full steam ahead. I think it will be several years before we see anything like true 4K *gaming* performance in the consoles. I mean, we dont even have Zen yet, and Zen APUs are even further in the future, with our without HBM, and we have no idea about performance or cost.

Eurogamer is more credible than the previous source, I think it's pretty safe to say this product actually exists. But who knows what kind of upgrades/enhancements Sony is planning. Could be a totally new APU or simply an improved version of the existing Liverpool including 4K capabilities. I personally doubt Zen and/or HBM will be in the mix, but we will see.

 

[MajinCry]

Having a more powerful console doesn't mean that the library will be fragmented; just allow the graphical settings to be toggleable, a la PC.

 

[Headfoot]

Benchlife mentioned something about 256MB and a decoder chip. Could simply be bluray ready PS4 with the external VR part integrated as well.

I cant imagine they are going to fracture the PS4 segment gaming wise. Because that's the whole point of a console.

I agree with this. Just a PS4 with all the addons stuffed into the main box, but you could get basically the same thing for your existing PS4 by buying the separate boxes too.

Though I think there's a small chance we see an upclocked, die-shrunk version that will output 4k30 with it being mostly upscaled but with certain certified titles increasing the internal viewport to something above 1080 but far below 4k. Then UIs / 2d would look 4k sharp without dramatically increasing the actual 3d horsepower needed. The PS4 is clocked pretty low for power reasons so they could theoretically upclock it on 14 FinFET and stay easily within the power budget. All you'd need is extra Q&A time to ensure the new clock is good

 

[exar333]

Reason?

Because PS4/Xbone struggle with modern games at 1080P (which is 10-year old flat-panel display tech). They are slow.

In another year or two, a Steambox will wipe the floor with these consoles and provide all the features people want for their new 4K TVs.

 

[NTMBK]

In another year or two, a Steambox will wipe the floor with these consoles and provide all the features people want for their new 4K TVs.

Yeah, I'm sure people will love getting 3rd rate ports 3 years late.

 

[exar333]

Yeah, I'm sure people will love getting 3rd rate ports 3 years late.

Vs. already buggy 'AAA' titles that are impossible to mod on consoles? I enjoy both PC and console titles, but tell it like it is.

Take Fallout 4 for example. Only a fool would prefer that on console (poor visuals, slow, and unable to mod). 3 years late? Came out the same day!

 

[poofyhairguy]

Could simply be bluray ready PS4 with the external VR part integrated as well.

Yup, unless Sony is drinking crazy juice.

 

[Justinbaileyman]

Reason?

Because PS4/Xbone struggle with modern games at 1080P (which is 10-year old flat-panel display tech). They are slow.

In another year or two, a Steambox will wipe the floor with these consoles and provide all the features people want for their new 4K TVs.

My thought as well..

 

[el etro]

A nice idea:

Sony launches the project in 2017 as PS5, project uses Polaris 10 + Zen with GDDR5;

Sony makes the PS4 and PS5 intercompatible for a time, extending PS4 lifecycle. Games run at 1080p on PS4 and 4K at PS5;


After some years, Sony stop producing PS4.

 

[Rakehellion]

 

[Sweepr]

 

[NTMBK]

 

[frozentundra123456]

Those *were* the good old days!!!

 

[superstition]

4K is so silly.

The entire industry messed up.

HDTV should have been 720p then 1440p. 1080 should have never existed. 4K exists because it's easy for manufacturers of 1080 panels to make them, not because the 4K resolution is something particularly special.

1440 is plenty for HDTV, given normal viewing distances. For console gaming it would also be fine.

People should have focused on important factors, like contrast ratio, input lag (a huge problem with HDTVs), color gamut, and the like — instead of obsessing about pixel density. 1440 offers plenty of density considering the requirements for pushing very high-quality graphics at that size.

Until the people and environments start looking like real life then 1440 is plenty. We don't need low-grade graphics at 4K. Seeing a pixel is not as much of an issue as the quality of the gameplay and graphical realism/complexity.

4K is OK for monitors because people sit closer but it's overkill for console gaming. But, since we were fed 1080, a stupid standard, we're locked into the 4K overkill. As a result, no one is very happy with the state of console gaming. They're currently too weak for 1080 and getting to 4K is going to cost too much.

 

[superstition]

As for the cost of bringing HBM to consoles, one potential barrier is the cost of the bumped surface that's used. It may still be the case that only one OEM produces those surfaces. I read an article some time ago that said others are going to get into the game of making them but until economy of scale from that and better yields are available HBM will carry a premium over GDDR5.

If the cost can be dropped on the bumped surface, though, through economy of scale and better yields, then HBM might be able to be more cost-effective due to the reduction in board complexity and such.

 

[Exophase]

Am I doing it right?

 

[MrTeal]

Vs. already buggy 'AAA' titles that are impossible to mod on consoles? I enjoy both PC and console titles, but tell it like it is.

Take Fallout 4 for example. Only a fool would prefer that on console (poor visuals, slow, and unable to mod). 3 years late? Came out the same day!

I think you might be vastly overstating the number of people who mod Fallout 4 vs the number of people who bought it. Even if you could build a PC for the $350 you'd spend on a PS4 that's more powerful (and even that's a big stretch), the number of people people clamouring for moddable games is likely a lot smaller than the number just wanting something to plug into their TV and turn on.