In a way game engines give you a sandbox for making your game in. And they exist in the sandbox that is your computers operating system (OS).
Could game engines be more powerful if they were closer to being an operating system as in higher privileges and closer to the metal control.
We are seeing closer to the metal control of the GPU in Vulkan and DirectX 12.
Could or should game engines be given closer to the metal control of CPU’s and OS privileges.
Or what if Unity brought out a light super fast OS version of it’s engine designed to run faster and smoother than the default operating system.
This is after all the domain of dedicated game consoles that often outperform similar specification PC hardware simply because they run closer to the hardware on a dedicated game OS…
Or could existing OS’s do more for game engines by giving them more power or a dedicated game mode sandbox to run in?
It has little to do with running closer to the hardware. Xbox runs on windows basically. It’s about hardware dedicated for a specific task vs general purpose.
The purpose of OS is to provide abstraction layer and universal api for underlying hardware, babysit threads and filehandles, manage memory, etc, etc, etc.
Comlexity of OS is at least order of magnitude higher (or several) than a complexity of a game engine, which is the reason why sony used FreeBSD for PS4 OS instead of trying to “create their own casino”.
Some numbers:
Estimated costs of creating Linux kernel from scratch is 1 billion usd.
Estimated costs of creating Linux OS from scratch is ten billion usd.
Net income of Ubisoft is mere 500 million per year in 2016 - not enough to tackle this kind of project.
So I suggest to get back to earth from the clouds.
Even more silly, you do realize that the game itself has to run on windows or some other existing OS people have right? You can’t seriously think people will install a new OS, and reboot their computer to play your game?
Because there were obvious bottlenecks with the previous way of doing things and it was practical to offer improvements to the situation without having to reinvent millions of wheels.
Perhaps if BeOS hadnt died or SteamOS appeared to have any serious momentum behind it, it would be slightly easier to take alternative desktop OS ideas at all seriously in 2017. Even then, a proper description of the serious performance problem existing OS’s have, compelling technical detail on how this amazing new OS overcomes the problem, with some exciting demos of stuff only possible with the new OS would be the bare minimum required to sustain any interest.
This used to be the way things worked several decades ago. Games would run directly on hardware, with no intermediate layer. It’s technically still possible, but highly unlikely anyone will go this route.
In practical terms it’s considerably easier to build a game in an ill fitting OS then it is to build an OS from scratch. A bad OS is dramatically better then none at all.
In a way it’s why we all use Unity. It’s not perfect for any of us. But it’s considerably better then nothing.
I don’t like how microsoft thinks it’s OK to put adverts and paid services inside the actual OS install, so I can’t avoid it. I can get rid of it, but not without seeing it.