Vulkan 1.0.5 API Specification Released

[computerquip]

Yeah, that. It is indeed quite readable. It clearly says that, instead of writing a shader in SPIR-V (which is low-level and explicit, easy to consume by a driver), a user can write a shader in GLSL (which is high-level and complex), and the Nvidia driver will magically compile it in an efficient manner.

It says absolutely nothing about being efficient. It doesn't even say how it will compile the GLSL. In concept, it doesn't have too.

[M1kkko]

Don't think so: https://developer.nvidia.com/vknvglslshader

Yeah, that. It is indeed quite readable. It clearly says that, instead of writing a shader in SPIR-V (which is low-level and explicit, easy to consume by a driver), a user can write a shader in GLSL (which is high-level and complex), and the Nvidia driver will magically compile it in an efficient manner.

[dragorth]

Why does AMD invest in R&D, if they end either benefiting the competition or shooting themselves in the foot...

Because, according to them, they believe in an open game market.

[asdfblah]

Why does AMD invest in R&D, if they end either benefiting the competition or shooting themselves in the foot...

[ptrwis]

Only standarised features should go to common specification, vendors' extensions should be documented and maintained in their own ext-spec.

[computerquip]

Your link requires some authentication to open?

Don't think so: https://developer.nvidia.com/vknvglslshader

[M1kkko]

Really? You know, contrary to popular belief, the specifications are generally quite readable. Why don't you read it before spouting more filth?
http://developer.download.nvidia.com...lsl_shader.txt

Your link requires some authentication to open?

[computerquip]

There is a specification AND the Nvidia's proprietary driver, which happens to be the implementation that developers will actually be testing their software against. And the specification seems to be too high-level to accurately tell how Nvidia compiles the GLSL shaders to SPIR-V in their driver. There's so much newly-introduced complexity that even in the best case scenario, at least some difference in performance across implementations is to be expected.



Yeah, and the performance of the open source drivers has been in such a great shape for all these years.

Really? You know, contrary to popular belief, the specifications are generally quite readable. Why don't you read it before spouting more filth?
http://developer.download.nvidia.com...lsl_shader.txt

EDIT: Performance isn't an issue. Vulkan was designed to remove overhead and give application developers control, not to speed up implementation. The specification will never take into account "speed" which is a relative term anyways...

[microcode]

Honestly I think VK_NV_glsl_shader is a pretty dumb extension, but if you're porting a legacy GL application I can see it being helpful. To be clear for all those not paying full attention: they are merely documenting the extension, it is not in core Vulkan, and I doubt other vendors are interested in implementing it.

If you want to transition to Vulkan but retain your glsl shaders, I think you could fairly simply introduce Glslang; though honestly, Glslang is in many ways not as faithful an implementation of GLSL and a mature OpenGL driver is, especially when it comes to generating SPIR-V code.

I hope NVIDIA deprecates this, because the mere presence of this extension has the potential to degrade the quality of Vulkan driver implementations, leading to the same sorts of issues which prompted the spec to begin with.

[M1kkko]

Because there's no "proprietary implementation" to match. There's a specification.

There is a specification AND the Nvidia's proprietary driver, which happens to be the implementation that developers will actually be testing their software against. And the specification seems to be too high-level to accurately tell how Nvidia compiles the GLSL shaders to SPIR-V in their driver. There's so much newly-introduced complexity that even in the best case scenario, at least some difference in performance across implementations is to be expected.

EDIT: And really, the implementation is mostly done for all drivers. They've done it for over a decade in their OpenGL drivers. Alternatively, they can just use glslang to go from GLSL to Spir-V to their intermediary.

Yeah, and the performance of the open source drivers has been in such a great shape for all these years.