Originally Posted by MonarchX
1. I think people do not take into consideration how much more difficult it is to implement low-level API utilization for game developers, who are already hammered with overly short deadlines for more and more complex and graphically advanced products (games). Being able to optimize for so many possible PC configurations with a high-level API, so imagine how much time its going to take to optimize for all those configurations with low-level API.
2. Look at Steam, the biggest Digital Video Game Media distributor, and tel me how many high-end hardware-demanding indie games you see vs. how many AAA games you see on Popular or Upcoming games? I think AAA games have higher sales per game (Steam's Top Sales), but there are fewer of them. Why complicate indie-game development with more complex low-level API when its not necessary for their simple-graphics needs?
3. Adoption. Neither DirectX 12 nor Vulkan will become main-stream this year. It makes more sense to slowly work on supporting low-level thoroughly while focusing on good-old DirectX 11, which still has at least some room for optimization.
4. Consoles. Regardless of what any game developer company says about their PC priorities, PC game support will always be the last item on their lists. Aren't they going to introduce DirectX 12 features into consoles? That is when it will make sensen to focus on DirectX 12 and Vulkan for better console-to-PC game porting.
5. They just don't care that much because this is not going to affect their video-card sales in the near future...
What we need is Vulkan renderer support in upcoming Doom game!
From my viewpoint - the option to do something on low level is good - an indie dev with his "simple gfx needs" does not need to go to low level to get the performance he desires. However, he has an option
to write a low level snippet of code that does something exactly like he needs/wants if it would simplify his work. For example - as far as I understand it might be hard to use the GFX card simultaneously for compute and gfx tasks under the current API's. Not impossible but hard. Under Vulcan a dev could write a custom function at relatively low level (but still sufficiently abstracted to work on different cards) that, for example, helps him/her to better simulate flow of liquids / gases / smoke / fire, etc for some kind of neat special effect in his indie game. After all, it's not the AAA games that innovate but most of the time fresh ideas and effects emerge from the indie games. In AAA games you get "it's a dog!!11!!" and "now the fish swims around!!11!!".