Precision

Cases

• case none

  No precision is specified, the GPU driver is allowed to use whatever level of precision it chooses. This is the default option and is equivalent to using dFdx() or dFdy() in text shaders.

• case coarse

  The derivative will be calculated using the current fragment’s neighbors (which may not include the current fragment). This tends to be faster than using .fine, but may not be suitable when more precision is needed. This is equivalent to using dFdxCoarse() or dFdyCoarse() in text shaders.

• case fine

  The derivative will be calculated using the current fragment and its immediate neighbors. This tends to be slower than using .coarse, but may be necessary when more precision is needed. This is equivalent to using dFdxFine() or dFdyFine() in text shaders.

• case max

  Represents the size of the Precision enum.