By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Sony - PS4.5/PS4K Is Codenamed 'NEO' And More Info

ZahaDoom said:

A ps4k isn't going to increase the resolution of the PSVR, it would only create two classes of VR users within the ps4 environment,  NEO'S and non Neo's with ZERO of the increased ability going to resolution and all the the extra hardware to push greater fidelity or framerate. 

And resolution isn't everything despite being most easily numerically compared... FPS and fidelity are important parts of game experience.

I'm spending enough on the PSVR as it is, now to be told I'm going to need a new console too? Huge mistake, and horrible timing with PSVR releasing.

Well, considering x-(half)-gen parity clause, you AREN'T being told that, at least at this point, are you?

Demonstrating the PSVR in store connected to a ps4k would be a big mistake, because owners then of ps4 might be reluctant to buy thinking their console can't handle it.

So... are PC games promoted based on min-spec, sub-market which publishers actually do expect to achieve sales with?



Around the Network
ZahaDoom said:
potato_hamster said:

That's the day Sony loses me as a customer if they pull crap like this. I'm not buying a new console every 3 years just to ensure that I get the version of the game the developers put most of their effort into.

 I hear you, I'm not sure why ANY gamer would want this , but there's a whole bunch of them here in this thread that seem to. 

 

 I just want to point to the mistakes Microsoft made launching the Xbox one, I'm sure they didn't think they were making any mistakes either but history proves otherwise.

 This is another move by someone in corporate telling people what they want instead of listening to what they want.

Sony is only a decade removed from the botched launch of the PS3. It humbled them.  They promised they would listen to their fans, and make choices to earn back their fans trust. Slowly, but surely it worked, and the PS3 eventually surpassed the X360 in sales. Enter the PS4. MS announces the X1 to mass hysteria, the idea is stillborn, and Sony capitalizes on it by bullet point after bullet point that screamed "we're listening and we care about what you want and we're showing it".

That was just three years ago.

And what do we have now? If this is true Sony has let the sale of 40 million PS4 go to their heads and are pissing off fans and developers alike with this horrible idea. It's like they have totally forgotten them what got them to 40 million consoles sold to begin with. Now I'm not buying another PS4 game until after E3 when we see what's what. The PS4 and it's games, controller, camera, and headsets might be sold for a new graphics card for my PC come June.



potato_hamster said:
Pemalite said:

The game engines already support the hardware. 

It is literally a matter of tweaking sliders as API's, drivers, OS everything else will remain the same.

If anything Multiplatforms will be closer to the PC now in regards to graphics as most PS4 games were only high-graphics compared to the PC's Ultra.


The game engine is optimized for a specific hardware configuration. it is optimized for bottlenecks specific to that hardware configuration. You just can't take the existing hardware engine and say "we'll just tweak this number and that number and this other number" and boom an engine just as optimized for PS4K and it was for PS4.  ... The PS4K build will have to be tested as equally as the PS4 build is now. That's a fact.

But x-gen PS4.0/4.5 games DON'T have to be "just as optimized" for PS4.5, they can take advantage of the X% more CPU and memory bandwidth and Y% more GPU power to add/boost performance.  PS4.5 will have a significant amount of free capacity beyond PS4, and you CAN just turn up sliders for res, FPS, lighting to fill that.  You don't have test equally because the architecture IS largely just scaled up.  YES, to "fully optimize" for PS4k would be more work, but you don't have to "fully optimize" in order to get a product that has noticeable benefit from running on PS4.5.  

Pemalite said:

For instance... Whilst the CU's have doubled, everything else hasn't... Suddenly the system just gained a bunch of new bottlenecks that are different or more pronounced from the original machine.

If CPU/Memory BW has been boosted by 25% while GPU doubled, it's safe to assume 25% headroom for scaling without running into different bottleneck.
To FULLY leverage the new (half) gen platform, will take more work, but there is plenty of scope for "easy" gains with little dev/testing... IMHO.



The industry is general is now going to be changed I think because of this. Iterative console refreshes will become common place for better or worse.



Soundwave said:
The industry is general is now going to be changed I think because of this. Iterative console refreshes will become common place for better or worse.

So basically buying psvr in October is a bad idea, as most likely a 1440p neo psvr headset will follow in a year or 2 to take advantage of the extra resolution the neo can provide. And the original one will be cheap second hand. Certainly not worth 399 now anymore.



Around the Network
potato_hamster said:
Pemalite said:

 

 

 

The game engines already support the hardware.
It is literally a matter of tweaking sliders as API's, drivers, OS everything else will remain the same.

If anything Multiplatforms will be closer to the PC now in regards to graphics as most PS4 games were only high-graphics compared to the PC's Ultra.


The game engine is optimized for a specific hardware configuration. it is optimized for bottlenecks specific to that hardware configuration. You just can't take the existing hardware engine and say "we'll just tweak this number and that number and this other number" and boom an engine just as optimized for PS4K and it was for PS4. As you've mentioned, different hardware configurations introduce different bottlenecks. These have to be accounted for, and solutions need to be developed. While you're right that it's less work than say optimizing an engine for the Nintendo NX, it's still a significant amount of work. People are going to spend months developing it, and engine revisions and improvements will take longer going forward. And again, there's literally doubling the amount of QA work needed. There's no two ways around that. The PS4K build will have to be tested as equally as the PS4 build is now. That's a fact.

This will make games more expensive to develop for PS4 going forward. Period. That means if developers have the option, they won''t support it, and if they don't they'll cut corners in other ways to keep the development costs the same. Budgets only increase if expected sales increase, and this specification bump gives no reason for developers to expect additional sales.

Or. Games will just be developed with the PS4 4K and PC in mind (Ultra graphics) and continue the practice of downgrading the games for the regular PS4. (High equivalent graphics.)

Nate4Drake said:

Nope ! Sony and AMD have a much higher understanding than anybody here about hardware, and 218GB/s memory bandwidth is more than enough considering The new Polaris Architecture is a lot more efficient, and "less memory bandwidth hungry"; also, don't forget about PS4 being heavily geared towards asynchronous compute which will allow non-graphics calulations to be done on the GPU !!! Plus, some CUs can be used enterely for non-graphics calulations, working together with the CPU, increasing massively the overal "non-graphics calulations" capability !!!  Don't say things you have no idea, please. Let the experts work on it. PS4 NEO will be as balanced as PS4, just MORE powerful.

1) We have no idea if the PS4 4K will use Polaris.

2) If you think 218GB/s is enough bandwidth for 4k AAA games, then... You are highly mistaken, I am going to assume you have never used high-end hardware and thus never gamed at resolutions that exceed 1080P... So allow me to educate you on this... Higher resolutions need allot of bandwidth.

3) Some non-graphics calculations are suited to being executed upon the GPU, some are not. - Because some calculations are very single-thread heavy, which a GPU is not adept at due to their highly parallel nature. There is a reason why CPU's and GPU's are still seperated functionally, because a CPU is adept as some tasks, the GPU is adept at others, both can do a "little" bit of each others work.

HoloDust said:
Pemalite said:

The PS4's GPU is close to a Radeon 7850.
This would put it near the Radeon 7970... At-least in terms of CU's.
We have no idea if it's using GCN 1.3 with it's colour compression or not, how many Texture Mapping Units or Render Output Pipelines or even if they bolstered the geometry units.
Way to many unknowns.

28nm is perfectly feasible, they will likely be aggressive with the voltages.
As a fabrication process matures you can tweak things for lower power consumption and more performance.

Performance wise, sure, numbers on the sheet are pretty similar to 7850. But it's 1280:80:32 chip with 2CUs disabled...I just don't know (nor was I ever able to find any info) is it Pitcairn XT with ACEs from Hawaii or is it Hawaii Pro cut in half (with 2 CUs disabled).

36CUs amount to 2304 shaders, which honestly sounds a lot like 40CUs (2560 shaders) of 290/390 cut down by 4CUs for better yields...so maybe 2304:144:64...which seems a lot to put onto 28nm if they want to keep TDP low.

I guess 28nm might be feasible if they're ditching PS4's design and going more toward the original PS3 TDP route...which I wouldn't be surprised honestly, if anything I was surprised by low TDP design of PS4 to begin with.

Would be pretty funny if AMD changed the amount of ALU's in a SIMD and the amount of SIMD's in a CU for the PS4 4k.

Need more info on the chip though.
Even with 2304:144:64 core layout, it would still only be a Radeon 7970 in terms of performance due to clocks, Which in my opinion, is what the PS4 should have released with.

Pretty sure the PS4's GPU is just an enhanced Pitcairn, GCN 1 chip.

The PS4 4k might not even use Hawaii, it *could* use a pair of Pitcairn dies.
AMD's GCN layout is also fairly modular, so it's fairly easy to scale it up and down as needed.




www.youtube.com/@Pemalite

potato_hamster said:
AnthonyW86 said:

You are forgetting the people who will upgrade to PS4K, they will sell their current PS4 off to other people. So every single PS4K sold increases the install base(regardless of already owning a PS4 or not), aside from sales of the original PS4 system wich will likely go down to $299.

This is a mid-gen upgrade, so the PS4 will get software support for years. That means that every PS4 out there is part of the active user base, unlike many PS3's collecting dust shortly after PS4 released. PS4 and PS4K add together to create one large install base. And not just for games, also for PS Plus.

Or they keep their existing PS4 and just buy a new PS4K.

It's still irrelevant, you haven't tied how the PS4K increases the PS4 userbase even larger than the PS4 would have without it. You can't just simply look at the PS4K numbers sold and think "well there's your answer" because 90-95% of them might have bought or kept their PS4 otherwise, or the people that bought thier used PS4 planned on buying a PS4 anyways and was just waiting for another price drop.

Most third party developers are going to look at the PS4K, and just as they decided with the  Sega CD, Sega 32x, N64 expansion pak, PSP hardware revisions, and New 3DS, they're going to think "not worth it", and just make PS4 games.

This is a waste of time and money for everyone involved. This is so astroundingly dumb. I really thought Sony wasn't that stupid.

Come back again in a year after the holiday season and see if you still feel the same way when Sony has announced record console sales and PS Plus subscription numbers.



Pemalite said:
potato_hamster said:


The game engine is optimized for a specific hardware configuration. it is optimized for bottlenecks specific to that hardware configuration. You just can't take the existing hardware engine and say "we'll just tweak this number and that number and this other number" and boom an engine just as optimized for PS4K and it was for PS4. As you've mentioned, different hardware configurations introduce different bottlenecks. These have to be accounted for, and solutions need to be developed. While you're right that it's less work than say optimizing an engine for the Nintendo NX, it's still a significant amount of work. People are going to spend months developing it, and engine revisions and improvements will take longer going forward. And again, there's literally doubling the amount of QA work needed. There's no two ways around that. The PS4K build will have to be tested as equally as the PS4 build is now. That's a fact.

This will make games more expensive to develop for PS4 going forward. Period. That means if developers have the option, they won''t support it, and if they don't they'll cut corners in other ways to keep the development costs the same. Budgets only increase if expected sales increase, and this specification bump gives no reason for developers to expect additional sales.

Or. Games will just be developed with the PS4 4K and PC in mind (Ultra graphics) and continue the practice of downgrading the games for the regular PS4. (High equivalent graphics.)

Nate4Drake said:

Nope ! Sony and AMD have a much higher understanding than anybody here about hardware, and 218GB/s memory bandwidth is more than enough considering The new Polaris Architecture is a lot more efficient, and "less memory bandwidth hungry"; also, don't forget about PS4 being heavily geared towards asynchronous compute which will allow non-graphics calulations to be done on the GPU !!! Plus, some CUs can be used enterely for non-graphics calulations, working together with the CPU, increasing massively the overal "non-graphics calulations" capability !!!  Don't say things you have no idea, please. Let the experts work on it. PS4 NEO will be as balanced as PS4, just MORE powerful.

1) We have no idea if the PS4 4K will use Polaris.

2) If you think 218GB/s is enough bandwidth for 4k AAA games, then... You are highly mistaken, I am going to assume you have never used high-end hardware and thus never gamed at resolutions that exceed 1080P... So allow me to educate you on this... Higher resolutions need allot of bandwidth.

3) Some non-graphics calculations are suited to being executed upon the GPU, some are not. - Because some calculations are very single-thread heavy, which a GPU is not adept at due to their highly parallel nature. There is a reason why CPU's and GPU's are still seperated functionally, because a CPU is adept as some tasks, the GPU is adept at others, both can do a "little" bit of each others work.

This isn't for 4K gaming, I think the term "PS4K" just got thrown around mistakenly probably because it does allow for 4K Blu-Ray movie playback. PS4.5 is more accurate, it's a huge GPU upgrade with a moderate CPU upgrade. 



mutantsushi said:
potato_hamster said:


The game engine is optimized for a specific hardware configuration. it is optimized for bottlenecks specific to that hardware configuration. You just can't take the existing hardware engine and say "we'll just tweak this number and that number and this other number" and boom an engine just as optimized for PS4K and it was for PS4.  ... The PS4K build will have to be tested as equally as the PS4 build is now. That's a fact.

But x-gen PS4.0/4.5 games DON'T have to be "just as optimized" for PS4.5, they can take advantage of the X% more CPU and memory bandwidth and Y% more GPU power to add/boost performance.  PS4.5 will have a significant amount of free capacity beyond PS4, and you CAN just turn up sliders for res, FPS, lighting to fill that.  You don't have test equally because the architecture IS largely just scaled up.  YES, to "fully optimize" for PS4k would be more work, but you don't have to "fully optimize" in order to get a product that has noticeable benefit from running on PS4.5.  

If CPU/Memory BW has been boosted by 25% while GPU doubled, it's safe to assume 25% headroom for scaling without running into different bottleneck.
To FULLY leverage the new (half) gen platform, will take more work, but there is plenty of scope for "easy" gains with little dev/testing... IMHO.

You actually do have to test equally. What on earth are you talking about? You have to be able to demonstrate to Sony that you have throughly tested your game, and even then they don't believe you and test it for you. You cannot assume your game will "just work" on the PS4K. You have to demonstrate that it does, that your enhancements, that the memory allocations you made still result in a game that doesn't hard or softlock. Because if Sony finds one of those while certifying, kiss your application money goodbye and start again. That is how it works. Want to go lazy on testing and have to go through certification three times (been there!) then that's tens of thousands of dollars you just lit up in flames. The same goes for other hardware manufacturers. Even though I've never worked on a New 3DS games, I do have colleagues that have and they literally doubled their alloted QA time and budget to support the new specification, because Nintendo has similar demands if you want to get a game on their system.

Ans I'm not sure why you think BOLDING AND CAPITALIZING YOUR TEXT MAKES YOU MORE RIGHT but you're still wrong about the whole "slider" bit. If the PS4K eliminates bottlenecks the PS4 has, and your engine has been designed around creating specific timing or ordering or instructions to minimize the impact of that bottleneck then you have to create an entirely new solution for the PS4K. And while it's fine and dandy to say "well it's just larger allocations of resources to handle" it's entirely possible that those larger resources are now fetching data from caches at a rate that's faster than your engine was designed to handle, meaning you now have to account for a potential backlog of instructions which could lead to further bottlenecks. My point is -  you still have to optimize for the new hardware! You always do with consoles. It's not just a matter of increasing memory allocations and telling the engine the CPU clock speed is 1.5x larger and mailing it in.



SvennoJ said:
Soundwave said:
The industry is general is now going to be changed I think because of this. Iterative console refreshes will become common place for better or worse.

So basically buying psvr in October is a bad idea, as most likely a 1440p neo psvr headset will follow in a year or 2 to take advantage of the extra resolution the neo can provide. And the original one will be cheap second hand. Certainly not worth 399 now anymore.

Not sure why you needed to quote Soundwave there to provide context for your post...

But I disagree with you, IMHO the increased power of PS4.5 will just be used for increased visual fidelity, fps, lighting, 
PS4.0 VR will skimp on those, and PS4.5 will enable them, that's it.
PS4.5 is not so powerful they can push high res AND those features. (only +25% CPU/Memory BW)

Look at this way: Morpheus is launching at the SAME TIME as NEO, Sony wants to create a platform around it, so why fracture base like that?
Why wouldn't Morpheus have the SAME life span as PS4.5 when they are launched at the SAME time?
As is, they sell PS4.5 as premier Morpheus experience from Day 1, while also being compatable with PS4.0.
If they are going to launch higher spec VR platform, it will be with PS5 down the line a bit where there will be more reason/benefit for doing so.
You are approaching things as if Morpheus VR "Belongs to" PS4.0, but it doesn't, it will be co-launching alongside PS4.5.