Aura7541 said:

I can imagine developers releasing patches so that their games can be compatible with the Scorpio, if the hypothetical situation is true. Even then, that takes a lot more work than implementing a slider.

Microsoft hasn't mentioned MAUs over the last few NPD statements. While I agree that the XB1 currently has an online userbase problem, I think the homogeneity of the console's library is the even bigger problem. A monotonous library will lead to a monotonous userbase, which will limit the appeal of games of certain genres like JRPGs.

Yeah, I was aware of the problems the design of the XB1 would engender but because I'm stupid when it comes to tech I didn't know if there would be a simple work around. To be honest, I'm still not convinced the Scorpio really exists and think it could be simply a rumour leaked to stop more 360/XB1 owners buying a PS4/Neo. The biggest problem MS have is the online userbase, which of course is why they're now suddenly keen on having crossplay with the PS4. People gravitate towards the console their friends play games on and online is at the centre of that. Make it crossplay and that's no longer a problem.

Mark Cerny was the gift to Sony that keeps on giving, I don't know if he should get the full credit on this, but making sure the platform was as standard as it was (basically an X86 architecture easily upgradeable), made it impossible for backwards compatibility on one hand, but pretty much secured cheap and powerful hardware that should just carry an entire userbase over to the new consoles without having to reset to 0 everytime they implement an upgrade. The people you mentioned complaining at first are pretty much by now forced to get on with the program (since every other console manufacturer is releasing new hardware) and by the time the Neo is out they won't think too much about it. For people who bought it day one, probably enough time passed that they are ready for something shinier, and for those who bought a PS4 not that long ago they are still good with their system and won't be left behind.

For developers what's more attractive? Creating a game for a userbase of 50M (by the time Neo is out probably) or launch some exclusive on a new platform that may or may not perform well in the long run?

Sony might have the chance to lock their userbase in with this move, giving them all that PC Steam users have enjoyed for years and years. Upgrading your machine doesn't render your game library useless anymore, so what reason would those users have to switch sides?

We will have to see how it plays out, but I think there was some high level 3-dimensional chess when they decided for this strategy.

Bravo.

It's a problem MS have always had though. They push Gears, Halo and Forza so hard that their userbase is always centred around these game. Not only that but MS openly promote Christmas with these games which creates a lull for the rest of the year.

I don't think it's a matter of 'want', but a question of 'can'. Remember that the current XB1 games are using a DDR3 + ESRAM memory architecture that is itself taking space on the APU. The only way they could get some type of APU that would match or overtake the PS4 (neo or otherwise) would be by ditching the ESRAM and going for GDDR5 like the PS4 did. This would mean that you couldn't simply plug and play the same games onto the new box without them being worked around the new architecture. This would take time and $ to do. When people were worried about the PS4 and Neo split, that would be the equivalent of moving a graphic setting slider from mid setting (PS4) to high/ultra (PS4 Neo). For the new Xbox it would be more like porting a game from Xbox 360 to Xbox 1 without emulation.

I'm sure some people will bring some crazy ideas like cloud computing and secret spicy sauce soon enough to explain it away.

Mark Cerny was the gift to Sony that keeps on giving, I don't know if he should get the full credit on this, but making sure the platform was as standard as it was (basically an X86 architecture easily upgradeable), made it impossible for backwards compatibility on one hand, but pretty much secured cheap and powerful hardware that should just carry an entire userbase over to the new consoles without having to reset to 0 everytime they implement an upgrade. The people you mentioned complaining at first are pretty much by now forced to get on with the program (since every other console manufacturer is releasing new hardware) and by the time the Neo is out they won't think too much about it. For people who bought it day one, probably enough time passed that they are ready for something shinier, and for those who bought a PS4 not that long ago they are still good with their system and won't be left behind.

For developers what's more attractive? Creating a game for a userbase of 50M (by the time Neo is out probably) or launch some exclusive on a new platform that may or may not perform well in the long run?

Sony might have the chance to lock their userbase in with this move, giving them all that PC Steam users have enjoyed for years and years. Upgrading your machine doesn't render your game library useless anymore, so what reason would those users have to switch sides?

We will have to see how it plays out, but I think there was some high level 3-dimensional chess when they decided for this strategy.

Bravo.

Agreed. What keeps people locked into an ecosystem? Exclusives, and soon the XB1 won't have a single exclusive. I've always likened Sony's strategy to chess and Microsoft's strategy to poker. One is played with guile and patience while the other is played with money and bluffs.

You do realise you can have both eSRAM and GDDR5? The eSRAM could be used as an L4 cache for the CPU and be used for retaining backwards compatability. It could also be a seperate chip that isn't a part of the main SoC and can also be used to reduce power consumption.

Besides, the Xbox One's SoC is built at 28nm, the eSRAM takes up something like 1.5 billion transisters, at 28nm this is a massive 30% of the chip... But if you drop down to 14nm and blow out the chip size, it might only be taking up 10% of the chip or less.

Whilst I do agree that Cerny made the right decisions at the right time, backwards compatability can still be broken in the future.

And not only that but the PS4 *could* have offered backwards compatability with the PS3 via several methods... Like including the original PS3 SoC or a software approach like what Microsoft has done.
Though... I think having hardware based PS2 and PS1 backwards compatability would be better than PS3 IMHO, I wouldn't say no to a limited run of PS2 consoles though with HDMI support for up-to 4k upscaling.

Halo and Fable has kept me gravitating towards Xbox, Gold kinda' locked me in.
With Fable now gone...
If Halo went to PC though, I will likely keep the Xbox, it's a multimedia powerhouse, it's probably one of the best HTPC devices you can get, especially with Kinect, which kinda' undermines the Xbox brand as I would rather watch a movie on it than play a game right now. :P

This also means they re useing a cheap chip with high power effeciency.

At 14nm FinFET this chip clocked at 911mhz probably wont generate all that much heat.
The 1200mhz+ version of the RX 480 is probably around 135watts (with a 150w limit)
At 911mhz clock speed, it ll be under 100watts for the GPU.

The chasis for the PS4 might not have to change too much.
Basically the new PS4 neo could end up looking much like the older model (same small form factor).

The only bad thing about it is - I was planning to buy this GPU and be happy for years to come, but if in a matter of months it's going to be the go-to console GPU, it means PC will soon need something beefier... RX 480X minimum. At least I hope AMD GPUs will be perfectly optimized for games from now on.

Yes all that is true but we have to take 2 things into account, price and the current rumors.

There is no magic here, if MS wants a console that is as powerful, if not more powerful as the PS4 Neo within the same pricerange they need to remove ESRAM. If it's included in the same APU, it will reduce yeilds and space for CUs. If they have a different apu that somehow includes the same or more CUs as the Neo while remaining with ESRAM then for sure it will be a much more expensive APU. 

In a world where price was not a factor of course they could throw anything inside the console and have it serve coffee as well, but that's not how things work.

The same goes for the case of PS4 offering backwards compatibility in the way you are describing. It would not meet a competitive price point first of all, and more importantly PS3 users would still not be able to play PS4 games. Now as it stands original PS4 owners can play Neo games and probably will be able to play PS5 games (less demanding ones i would assume) at lower settings. All because of settling down on a non exotic PC architecture :)

Polaris isn't really a game changer in the PC landscape, that was never it's intention, not from a performance perspective anyway.

Vega is where AMD really offers a high-end chip.
Navi with it's next-gen memory and likely an explosion in die-size, should be where things really start to get interesting as far as performance goes.

As for AMD GPU optimizations in the PC landscape, won't happen.
nVidia and Intel control larger market shares in the PC landscape, nVidia works closely with game developers to implement nVidia-specific technology's and optimizations and advertises the fact.
AMD's drivers can also be hit or miss... And PC's have a different software ecosystem. (OS, API's, etc'.)

That assumes that Microsoft isn't happy enough to eat some of the cost and that assumes Sony isn't profiting from the Neo.
Both have happened historically with Consoles.

As for backwards compatability, let's assume Sony sticks with "non exotic PC Architecture". - What happens if some of the fixed function units were removed from the GPU that PS4 games rely on? You break backwards compatability. (And we have allot of fixed function units even in modern GPU's.)
This has happened multiple times historically.
For example, the Xbox 360's Tessellation unit rely's on N-Patches to perform tessellation, which is incompatible with the Xbox One and Playstation 4's Tessellation unit from a hardware perspective.
Or hows about when the TnL fixed function unit from GPU's was eventually removed? One has to assume that games on consoles aren't targeting high-level API's and are building closer to the metal where such things are more important than on PC, where things can be abstracted and thus compatability is less of a fickle thing.


What happens if some SIMD instructions are removed from the processor that games rely on? You break backwards compatability, this has happened before with 3D Now! MMX, etc'.

Console generations takes year as well. Going from the PS3 Geforce 7 GPU to todays Geforce '18' GPU's is a night and day difference architecturally, games built and targeting every single nuance would be incompatible, that is the same kind of jump between a PS3 and the Neo and one would assume would be a similar kind of jump between the PS4 and PS5.

it would be speculation behind speculation to now assume MS would eat the costs for lower yeilds of the apu to once again include esram. they need to pull the bandaid at some point. the same way Sony did when ditching the Cell and their own exotic architectures in favour of a PC type approach.

i think their custom api apparently very much Vulkan like (or Vulkan based even, not 100% sure on that) should minimize the differences when coding for the PS4 and the Neo, very much like like you can run the same game on PC on any gfx card of the last 10 or so years (not all games of course but the majority) at different graphical settings.

lets keep in mind that there will be a lot less variety to code for on the console side. if updates will come every 3/4 years i'm confident that, sure, there will be a break off point at which certain games stop working on older hardware, but it should remain a more smooth transition. kind of like you can go now on steam and play half life 1 on hardware that is in no way close to the one existing in 99 or 2000 or whenever it was that such game came out.