w component of SV_Position/gl_fragcoord

Hello.
In DirectX 11, w component of SV_Position in pixel shader is view space depth.

hmm, ok. that means if i want to convert dx11 shader that uses w (viewspace depth) to vulkan shader (gl_fragcord), i need to do 1/w to get correct result?

Fortunately, 1/w is already stored in z component of SV_Position in pixel shader.

Not sure what exactly you mean by "doing 1/w", but in both HLSL and GLSL, you do not need do divide your vertex shader output by w. In all APIs your shader output is in homogeneous 4D space, the projection to w=1 plane is done by the rasterizer after culling, clipping and other steps.

Be aware though that while D3D and Vulkan use normalized device Z coordinate range [0, 1], OpenGL by default uses [-1, +1] (unless you use extensions to fix this). This affects how z from NDC is transformed to depth and defines how you should set up your projection matrix.

For people who has the same question.