Vulkan vs DX12 compatibility?

[therealnickdanger]

Based upon what I have read so far:

DX12
1. Feature levels for different GPUs
2. Windows 10 only
3. GPU driver must support

Vulkan
1. Any GPU capable of OpenGL 4.0 or OpenGL ES 3.1
2. Any operating system
3. GPU driver must support

On paper, why would anyone want DX12? Xbone compatibility? There must be more to it than this, so please elaborate.

 

[ShintaiDK]

For the same reason we picked DX the last 20 years.

Ask yourself what Vulkan does, that whatever previous OpenGL couldn't do. And look on why OpenGL failed.

 

[NTMBK]

Based upon what I have read so far:

DX12
1. Feature levels for different GPUs
2. Windows 10 only
3. GPU driver must support

Vulkan
1. Any GPU capable of OpenGL 4.0 or OpenGL ES 3.1
2. Any operating system
3. GPU driver must support

On paper, why would anyone want DX12? Xbone compatibility? There must be more to it than this, so please elaborate.

Better developer tools. Better investment in drivers. Quicker to add new features (no getting stuck in Khronos committee for years).

 

[therealnickdanger]

For the same reason we picked DX the last 20 years. Ask yourself what Vulkan does, that whatever previous OpenGL couldn't do. And look on why OpenGL failed.

That's my suspicion as well.

Better developer tools. Better investment in drivers. Quicker to add new features (no getting stuck in Khronos committee for years).

Will adding features be as important this time around, given the amount of flexibility and control that Vulkan (and DX12) gives to developers? Likewise, shouldn't this help to alleviate some of the game-specific driver tweaks that GPU vendors need to provide?

 

[Dribble]

The dev tools will be key - all this lower level programming is going to be even trickier then gpu programming already was, and lead to even more bugs, the quality of the tools to fix these bugs will be key. You can bet DX12 will have much much better dev tools (MS make Visual Studio after all) then Vulkan which doesn't have any single company backing it in the same way and hence willing to make that investment.

 

[Beer4Me]

DX12 will always win because the games are on Windows. No one uses Linux to game. Where Vulkan could help is with consoles...

 

[poofyhairguy]

Where Vulkan could help is with mobile devices and ARM set top boxes...

Fixed.

 

[cytg111]

Based upon what I have read so far:

DX12
1. Feature levels for different GPUs
2. Windows 10 only
3. GPU driver must support

Vulkan
1. Any GPU capable of OpenGL 4.0 or OpenGL ES 3.1
2. Any operating system
3. GPU driver must support

On paper, why would anyone want DX12? Xbone compatibility? There must be more to it than this, so please elaborate.

It would be a win for everyone but microsoft, no doubt about it. Gaming is not locked into windows anymore, the tablet formfactor, consoles and Apple with osx/ios .. I am hoping for change but not holding my breath.
ONLY reason I would rock/dualboot windows10 would be for mocking about and gaming, mocking about I can mostly do in a VM.

 

[dogen1]

DX12 will always win because the games are on Windows. No one uses Linux to game. Where Vulkan could help is with consoles...

Consoles have no need for vulkan. It would be worse than what they use already.

 

[Headfoot]

I think the big deal will be DirectX12->Vulkan shims, not native Vulkan. Valve was going around talking up their DX11 to OpenGL wrapper for SteamOS/Linux and they've actually gotten a decent amount of games to use it. It loses a lot of performance going to Linux but less than earlier efforts had.

I bet dollars to donuts that Valve will upgrade to a DX12->Vulkan shim after Source 2 initially debuts or go Vulkan outright and continue their SteamOS wrapping efforts.

My guess is that due to the similarity and low level nature of both API's there will be less performance loss due to the wrapper/shim in DX12->Vulkan than DX11->OpenGL

 

[3DVagabond]

DX12 will always win because the games are on Windows. No one uses Linux to game. Where Vulkan could help is with consoles...

Vulcan runs on Windows. The kicker will be if it runs good enough to make it attractive over DX.

 

[Pottuvoi]

Better developer tools. Better investment in drivers. Quicker to add new features (no getting stuck in Khronos committee for years).

It could be that Vulkan reveals more features from the start than DX12, so it certainly will be interesting to see where it will go.
Also not sure what level of features will need to go trough Khronos and what can be exposed as simply allowing new buffer formats etc.

 

[NTMBK]

Fixed.

Apple have already chosen Metal over Vulkan. The only cross portable API is OpenGL ES.

 

[Dribble]

Where Vulkan could help is with mobile devices and ARM set top boxes...

Because mobile devs are really going to want to code to the metal for 100 different makes of device, and set top boxes really need hardcore low level GL to display a few menu's...