Vulkan-Docs VS vulkan-gp

Pipelines (or in general terms PSOs) are the most problematic aspect of Vulkan / DX12 - much more than synchronization! Large parts of the gamedev industry seems to recognize all the performance issues with pipelines and therefore companies are experimenting with newer models like the VK_EXT_shader_object extension ("Vulkan without Pipelines": https://www.khronos.org/blog/you-can-use-vulkan-without-pipe...).

I've written a detailed comment about this before here (https://news.ycombinator.com/item?id=37843946#37845431) but for a much more comprehensive explanation by an engineer from Nintendo read the initial proposal for the VK_EXT_shader_object extension: https://github.com/KhronosGroup/Vulkan-Docs/blob/main/propos...).

There's also Casey Muratori's mail to the Vulkan advisory on 2015 that basically predicts this whole clusterfuck would happen: https://github.com/cmuratori/misc/blob/main/vulkan_dynamic_s...

Apple -> MoltenVK is an emulation layer and doesn't give you as much control as using Metal directly.

Nintendo and Sony prefer their own APIs, NVN and LibGNM, and AIUI Vulkan is a second-class API on those platforms which does not offer as much power, it is widely understood studios use NVN and LibGNM to get access to the real hardware on those platforms.

Windows/AMD/NVidia/Intel -> HW manufacturers tend to prototype and release new features with D3D first and then 'backport' them to Vulkan after a while. DirectX 12 for example had mesh shaders for over 2 years before Vulkan got a vendor neutral extension for them[0]

Android and Linux are the only platform where Vulkan is a first-class citizen.

You could maybe argue Nvidia treats Vulkan as a first-class citizen because they tend to have vendor-specific Vulkan extensions for the latest features available before anyone else. But otherwise, no, Vulkan is not a first-class API anywhere except Linux and Android.

Graphics API wars are alive and well.

[0] https://github.com/KhronosGroup/Vulkan-Docs/issues/1423

However, this does not seem to work. For some reason, it seems like the very first frame submitted after the swapchain recreation never signals its fence. Here is the code if you would like to have a look -- swapchain recreation happens at line 1135, after calling vkAcquireNextImageKHR. I've left the log calls in to make it clearer what the output below means.