Some developers have reported that in order to keep the frame rate of the NEO version above that of the original PlayStation®4, the resolution of the display buffer must be reduced
It isn't false. But It's hilarious to think Ms releases a 3TF console end 2017.
That is true.Greenberg was very careful in NOT saying 'native'... There's a difference!!
If specs change at all from the leak to release, I think RAM bandwidth is the most reasonable change.Can someone shed light on the discrepancy between increased GPU and increased memory bandwidth? The former is 2.3x base but the latter is only 1.24x. Is this what could be contributing to the comment
If specs change at all from the leak to release, I think RAM bandwidth is the most reasonable change.
DunnoIs it possible Sony could secure GDDR5X for Neo?
No, not Xbox anywhere. Even if publishers doesn't support play anywhere (I do think they do, but that's a discussion for another topic) on Xbox from now on games are decoupled from hardware. This means that if a game supports both xbone and Xbox 2 it's the same game, the same online community. It just happens to run and/or look better on the newer console.
And BC doesn't solve that problem. If a game releases for both consoles they are most likely to not be interchangeable. So it's two games, with two different populations.
Ms universal platform was made with that in mind, not only the games, but the platform itself is the same, it will be the same network whether you are on xbone or xbox 2, playing the same games (bar those who are too demanding to run on xbone 1 which will come a few years later), with the same accessory support.
What I'm talking about is essentially what you already have on iOS or Pc, where you can upgrade your hardware and your games automatically run better, not because they are being emulated, but because the game itself was made to run and adapt to devices with different performances.
Yes I'm liking Microsoft's direction with the backwards and forwards compatibility, play anywhere, and monster specs in the Scorpio. I like knowing I'll always have my 360 and xbone games in my library. They've turned the ship around after their dismal start to the generation.From Microsoft's point of view it's probably a shame that the transition from consumers being being locked in to hardware, to locked into the store, is happening now. (After they lost a big lead to PS).
I won't buy a Scorpio because it's more powerful than a neo, I'll buy it because I already have a ton of Xbox games.
Similarly when it comes out a lot of PS owners will stick with neo because that's where all their games are.
Essentially I think the console market has got very sticky, as consumers will have a lot to lose if they switch brands (stores).
So my point is, that if ms want the Scorpio to convince PS owners to switch, it's going to have to be a very "cheap" price. That may not be their strategy however, they may just want to retain their existing customers and make another round of money out of them.
I'd love to see the numbers on this. Scorpio may be their last chance to significantly shift the needle and coax people to get "stuck" in their store. So how much is that worth?
Fake edit - I'm including the ever greater switch to digital in my assumptions.
3Tf is 2.3x the XBO GPUSoooo... people are actually trying to spin that Scorpio will actually be 3 TF? In 2017?
3 TF in 2017?
Soooo... people are actually trying to spin that Scorpio will actually be 3 TF? In 2017?
3 TF in 2017?
Except Scorpio launches 4 years after, not 3!3Tf is 2.3x the XBO GPU
Same as PS4 to Neo
Can't argue with that math!
OléGunner;210051951 said:Well I mean there will be a PS5 yeah!
But what can the generation jump truly achieve when all games will be shackled to PS4 Neo and potentially vanilla PS4?
Or maybe we're at that point in technology where old hardware like PS4/XB1 can still allow all devs to realise their game visions years from now.
I dunno.
Wait, it's taking MS an extra year to release a system that's actually 1.2TF weaker than the Neo and they are just lying about it to save face?
So could the Neo do games at 4K using this half-precision?
I do think it's pretty cool that Sony is actively encouraging devs to contact them for assistance if needed.You tell me
![]()
I do think it's pretty
cool that Sony is actively encouraging devs to contact them for assistance if needed.
I think Sony's up to the same tricks and the Neo is just a 2.1TFLOPS machine which is just a bit beefier slim PS4.It wouldn't be a lie & nothing was official about Neo as they was unveiling the Xbox Scorpio so they had noting to save face from.
What are you even talking about? I didn't say they will magically make everything 4K I said that they will use uprendering & they are.
no, I'm just saying he's taken a lot of crap in this thread but it doesn't phase him
I think Sony's up to the same tricks and the Neo is just a 2.1TFLOPS machine which is just a bit beefier slim PS4.
While I don't think MS is being misleading about the 6tf, there's a big difference between the info we have from MS and Sony. We have official documentation from Sony that has been leaked, with detailed specs that is giving the calculation of 4.2 tflops.I think Sony's up to the same tricks and the Neo is just a 2.1TFLOPS machine which is just a bit beefier slim PS4.
I feel like uprendering is a misleading term for reconstruction techniques. We're talking about the same things, just with different terms and everyone is getting confused."Uprendering" isn't a specific thing. The cheap rendering techniques you go on to mention in your next post are real, but they don't really fall under any specific category since there are a variety of ways to construct a higher resolution image out of lower resolution buffers.
I think Sony's up to the same tricks and the Neo is just a 2.1TFLOPS machine which is just a bit beefier slim PS4.
So the PS4 is 0.9 TF then?The problem with that is that Sony hasn't said the Tflops number & we no that it is 4.19tflops (single-precision ) by doing math using the CU count & clock speed.
By the way it is 2.1TFLOPS (double-precision)
Soooo... people are actually trying to spin that Scorpio will actually be 3 TF? In 2017?
3 TF in 2017?
Gotta go for that balanced system performance.A 3TF system with 320GB/s bandwidth and 12GB RAM. Seems legit.
A 3TF system with 320GB/s bandwidth and 12GB RAM. Seems legit.
A 3TF system with 320GB/s bandwidth and 12GB RAM. Seems legit.
A 1.3tflops system with 204GB/s bandwidth and 8GB RAM. Seems legit.
The standard measurement of FLOPs for GPUs, especially with gaming in mind is single-precision. If you see a FLOPs number for a GPU, it is single precision, unless specified otherwise. Half-precision and double-precision floating point ops are more often useful for other tasks. There could actually be efficiencies gained from using half-precision for certain tasks that don't require single-precision accuracy, but that's a side point.
Except that's not the case is it. It's actually 68GB/s DDR3 with a 32MB ESRAM 'scratchpad' having a max theoretical bandwidth of 204GB/s. But realistic performance of ~140 or so.
Your line of 'thinking' is interesting but very obviously wrong.
http://www.eurogamer.net/articles/digitalfoundry-vs-the-xbox-one-architects"The same discussion with ESRAM as well - the 204GB/s number that was presented at Hot Chips is taking known limitations of the logic around the ESRAM into account. You can't sustain writes for absolutely every single cycle. The writes is known to insert a bubble [a dead cycle] occasionally... one out of every eight cycles is a bubble so that's how you get the combined 204GB/s as the raw peak that we can really achieve over the ESRAM. And then if you say what can you achieve out of an application - we've measured about 140-150GB/s for ESRAM.
"That's real code running. That's not some diagnostic or some simulation case or something like that. That is real code that is running at that bandwidth. You can add that to the external memory and say that that probably achieves in similar conditions 50-55GB/s and add those two together you're getting in the order of 200GB/s across the main memory and internally."
So 140GB-150GB is a realistic target and DDR3 bandwidth can really be added on top?
"Yes. That's been measured."
How is it wrong? how you know the 320GB/s number isn't the same as the 204GB/s number?
They usually use the peak flops which would usually be the single-precision number on PC GPUs but mobile GPUs that use half-precision us the half-precision numbers . Polaris peak number is also the half-precision number.
Both have there plus and minus going for them .
The problem is the long jump is not as big as it use to be .
Which happen because of many factors .
It's not like PC devs target the highest end to begin with anyway .
There's actually something very unique about Ms platform : The fact that it's the exact same network, with the exact same APIs and libraries. You don't need to support cross play (save for the guidelines about handling different input methods), you just implement the network on your game and it works on Xbox, scorpio and pc if you like. Convenience is key for a feature to get developer support.You do know there's nothing unique about Microsofts platform in the respect to cross play . There's nothing stoping publishers from releasing crossgen games that play online together.
You say that is there a reason the same thing can't happen to ps4? The psn has matured enough it doesn't need another rewrite for next gen. It should be able to use the underlying code same as Microsoft. See what's playing and not what's they are playing on. With the neo coming out it proves that even a upgrade in specs you can using the underlying architecture of the network to play with ps4 players. Since the ps4 went x86 there is no reason to change going forward generations. So the stuff they have now will just work on better hardware. They no longer need to reinvent the wheel every time with their OS, tools and apps. the only thing the network needs to worry about when playing crossgen games with the ps5 is a text field.Well, the jump will not be as high if you buy every iteration, but if you skip some I don't see how the jump wouldn't be the same as a full generation.
This is actually what I like about iterative releases, it's ultimately your choice when to jump to new hardware and I think you'd be able to keep a console for longer than you would on a traditional generation cycle (a normal one, not the almost 10 year long the last one was)
There's actually something very unique about Ms platform : The fact that it's the exact same network, with the exact same APIs and libraries. You don't need to support cross play (save for the guidelines about handling different input methods), you just implement the network on your game and it works on Xbox, scorpio and pc if you like. Convenience is key for a feature to get developer support.
Which is also why we don't see much cross gen games with a shared community. The 360 live is different from the xbone's, you still have work to provide interop between those two. But from now on Xbox live is just a single entity on Xbone, Scorpio, pc and whatever comes next, so it's zero work for the dev to have those platforms sharing the same user base.
There's actually something very unique about Ms platform : The fact that it's the exact same network, with the exact same APIs and libraries. You don't need to support cross play (save for the guidelines about handling different input methods), you just implement the network on your game and it works on Xbox, scorpio and pc if you like. Convenience is key for a feature to get developer support.
Except that's not the case is it. It's actually 68GB/s DDR3 with a 32MB ESRAM 'scratchpad' having a max theoretical bandwidth of 204GB/s. But realistic performance of ~140 or so.
Your line of 'thinking' is interesting but very obviously wrong, for numerous reasons.
A 1.3tflops system with 204GB/s bandwidth and 8GB RAM. Seems legit.
How is it wrong? how you know the 320GB/s number isn't the same as the 204GB/s number?
http://www.eurogamer.net/articles/digitalfoundry-vs-the-xbox-one-architects
It kinda plays into the rumors that Neo only exists because the new chipset is actually cheaper to source than having to engineer the PS4 architecture into a slim would be.I think Sony's up to the same tricks and the Neo is just a 2.1TFLOPS machine which is just a bit beefier slim PS4.
I believe it. Curious to see if new regular ps4s are just under clocked neos. We going to have guides how to turn your new ps4 into a neo.It kinda plays into the rumors that Neo only exists because the new chipset is actually cheaper to source than having to engineer the PS4 architecture into a slim would be.
Yup. PC and Xbox dx12 are different code paths.Didn't the Doom dev, just kinda shoot this idea down? Saying that the DX12 API's for Xbox and PC are indeed different?
probably from this guyI believe it. Curious to see if new regular ps4s are just under clocked neos. We going to have guides how to turn your new ps4 into a neo.
Everythig you say is right . But you realy think MS is doing this, after announcing the most powerfull console ever Made thing?
At this point, announcing the end of MS console gaming would be the same.
New AI algorithms for peak performance New half-precision instructions deliver more than 21 teraflops of peak performance for deep learning.
Read more: http://wccftech.com/nvidia-pascal-gpu-gtc-2016/#ixzz4EXMPWVlX
Of course others can do too, but afaik ms is the only one providing the exact same network across different devices. And not just network, you can think of the bone dash becoming an ios like system, instead of having a different OS for each new console.You say that is there a reason the same thing can't happen to ps4? The psn has matured enough it doesn't need another rewrite for next gen. It should be able to use the underlying code same as Microsoft. See what's playing and not what's they are playing on. With the neo coming out it proves that even a upgrade in specs you can using the underlying architecture of the network to play with ps4 players. Since the ps4 went x86 there is no reason to change going forward generations. So the stuff they have now will just work on better hardware. They no longer need to reinvent the wheel every time with their OS, tools and apps. the only thing the network needs to worry about when playing crossgen games with the ps5 is a text field.
Dx12, and to a bigger extent even win 10 is a work in progress. Thought I think that even now much of the code would be shared already.Didn't the Doom dev, just kinda shoot this idea down? Saying that the DX12 API's for Xbox and PC are indeed different?
Sony don't particularly care if anyone buys the PS5. If we were still okay with buying games for our PS1s, that'd be more than okay with Sony. And supporting hardware ~3 years old doesn't seem to retard PC development to any great degree, so I think we'd be alright.That's nonsense. Why would the lack of forwards compatibility hurt them when the inclusion of it would slow down adoption rates of PS5 considerably and hurt creativity from developers who'd be bound to last gen hardware?
I'm pretty sure they never said anything of the sort.I think what he means is that developers should not be mandated to have to make sure every game on PS5 runs on PS4 even in the PS5 era(since NEO is a PS4, it would by definition mean PS4 would have to be supported to), as that would destroy the point of a PS5 bringing in new experiences and techniques that could not be done on the older hardware.
That's simply not true at all. If anything iterative releases bring new tech faster.
Look no further than the pc market, pc versions of last gen games already had the visual fidelity you would find early on this generation, only a few years prior.
Mobile phones are also another example on how games advanced when they moved to iterative releases compared to dedicated handhelds.
It's releasing a big jump every 6 or more years that brings advancements down, not to mention the impact it had on the industry when pc developers started supporting consoles and console only devs were left in the dust.