All of those, even the most fully featured OS don't need much bandwidth. The amount of bandwidth LPDDR chips provide is more than enough for any combination OS task/features you can name
It depends on how Sony engineers want to build the OS and the features they want to come at a zero cost to games if games want to use them and what features they want the system to handle at the same time as games (people did not seem to be happy about in game XMB on PS3
). The way they built the PS4 home screen had a lot of dynamic features and the technology chosen for part of it, and for the PSN Store app, requires a full browser running a JavaScript + WebGL webapp... see
https://siliconangle.com/2013/11/20/webgl-graphics-technology-powers-sony-playstation-4-ui/ and the comments toward the end of the article that the PS4 UI, especially on the home screen, is designed to be web developer friendly to speed up the design of pages, components like carousels, and forms and pop ups.
PS Vita, while much less powerful, ended up having an even more polished and fluid UI even though it did not have all the features PS4 launches with. PS4 UI probably allowed them to integrate with other third party or shared services using more standard (albeit less performant and much more resources hungry and complex to make them feel App/native like) components as well as to extend the UI with components generated by other teams (see WebComponents:
https://www.webcomponents.org/introduction#what-are-web-components-).
So flexibility and ability to get up and running faster... with the cost of extra resources on the CPU and memory size (check out your browser resources consumption on your PC to see what kind of monsters they are hehe) and...
Can you elaborate as i don't quite understand? Another reason i read on forums is that it would make the PS5 board more complex but then again ps4 already does it...
... there is an argument about HW complexity that would push to eliminate other secondary memory pools while keeping a big.LITTLE like approach so that power hungry CPU cores are not kept awake and switching when not demanding tasks (see background downloads) are running, but in this case the argument is perhaps not to make the problem worse by adding faster and more power hungry and complex to integrate memory to offload even more OS tasks from the primary processor cores to the secondary one (which will get a likely big speed bump too, expect dual core 64 bits ARMv8 cores possibly).
Also, if you could have enough RAM and CPU power to keep it all running (beside base security, networking, and general I/O features which makes most sense to isolate... see T1 and T2 chips in MacBook and iMac computers for example) on the same OS and memory space then you would not have to keep writing for and maintaining different code based with different tools, but you could do it all with the same code and tools you wrote the main OS, the developers’ SDK, the high level libraries, and the games with. There are some advantages to unified code bases too.
Maybe due to smartphones demand LPDDR chips are just as expensive as GDDR
Not as expensive, but certainly gets the cost closer and if you could drive the console board without adding an extra sizable pool of memory
Oh i agree, the biggest upgrade will be the CPU, a proper next gen jump, BUT the geek in me screams at the waste of not 1 but 2 zen cores, it just seems too wasteful