[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] WebVulkan

On Fri, Aug 5, 2016 at 6:06 PM, Ashley Gullen <ashley@scirra.com> wrote:
- improved performance due to reduced driver overhead
The way vulkan achieves those benefits (if properly used), is by changing the semantics of the API and putting the burden on the programmer to properly assemble a command buffer, and appropriate GPU code, to facilitate best use of the available GPU resources (of which there's a variety of different flavors, exposed by vulkan), and a vastly expanded repertoire of code and commands you can issue to a GPU.

You can't have the cake (a simple API) and eat it too (get the benefits without the changes).
I think it is entirely possible to get all these benefits in WebGL if browser makers implement WebGL with a Vulkan backend. There appears to be at least some interest in this in ANGLE: https://bugs.chromium.org/p/angleproject/issues/detail?id=1319
Implementing WebGL atop vulkan is interesting, but it's not going do any of what vulkan promises and simply serve as another means to work around buggy drivers. Which is worthwhile, but it's not the future. GL is the end of the line. If you don't come up with something, it relegates the hardware accelerated web to that 20 years out of date second class citizen, again.

We have time now, to think about this. But this time is finite. Stalling out on having no plan than WebGL isn't going to work.