Devices may indicate support for wave operations at run-time. #elif defined(UNITY_PLATFORM_SUPPORTS_WAVE_64) #if defined(UNITY_PLATFORM_SUPPORTS_WAVE_32) #if (( defined(UNITY_PLATFORM_SUPPORTS_WAVE_32) || defined(UNITY_PLATFORM_SUPPORTS_WAVE_64)) & ( defined(UNITY_COMPILER_DXC) || defined(SHADER_API_PSSL))) To simplify usage, we check for a supported compiler here before indicating that wave operations are supported. Shaders on these platforms may not always be compiled with a compiler that supports wave operations. Platforms may indicate support for wave operations at compile-time. Support for wave operations is intentionally limited to the compute shader stage in order to make this functionality available to a wider range of hardware.
If a function require to have both a half and a float version, then both need to be explicitly define Any function that require an explicit precision, use float or half qualifier, when the function can support both, it use real (see below) The function of the shader library are stateless, no uniform declare in it.
static float unpackedArray = (float)packedArray i.e don't use SetGlobalFloatArray or SetComputeFloatParams Rules: When doing an array for constant buffer variables, we always use float4 to avoid any packing issue, particularly between compute shader and pixel shaders This mean that these structure need to be padded. These structures need to be align on float4 to respect various packing rules from shader language. Structure definition that are share between C# and hlsl. Something that we can't control with the global namespace (uniforms get optimized out if not used, modifying the global CBuffer layout per kernel) The reason is that for compute shader we need to guarantee that the layout of CBs is consistent across kernels. All uniforms should be in contant buffer (nothing in the global namespace). headers from ShaderLibrary do not include "common.hlsl", this should be included in the. When declaring "out" argument of function, they are always last Do not use "in", only "out" or "inout" as califier, no "inline" keyword either, useless. uniform have _ as prefix + uppercase _LowercaseThenCamelCase constant floating number written as 1.0 (not 1, not 1.0f, not 1.0h) FragDefault / FragForward / FragDeferred use input/output as variable name when using these structures Input/Outputs structs in PascalCase and prefixed by entry type capital letter mean the vector is normalize, unless we put 'un' in front of it. use capital letter for regular vector, vector are always pointing outward the current pixel position (ready for lighting equation) Example: unL for unormalized light vector normalized direction are almost everywhere, we tag unormalized vector with un.
A space where the translation of the camera have already been substract in order to improve precision This is very important regarding physically based light unit and inverse square attenuation
#Unity palette swap using matrix code
The lighting code assume that 1 Unity unit (1uu) = 1 meters. For cubemap capture (reflection probe) view space is still left handed (cubemap convention) and the determinant is positive. Caution: When going from world space to view space, unity is right handed in view space and the determinant of the matrix is negative Unity is Y up and left handed in world space #pragma warning (disable : 3205) // conversion of larger type to smaller #if SHADER_API_MOBILE || SHADER_API_GLES3
0 kommentar(er)
