warp.Float

warp.Float

Generic annotation for any Warp floating-point scalar type.

Use Float to annotate parameters of generic Warp kernels and functions, either directly or as the dtype of a generic array, vector, or matrix. Float matches the following floating-point types:

• float16, float32, or float64.

• bfloat16.

• Python float, which Warp treats as float32.

Use Scalar to accept both integer and floating-point types.

For more on writing generic kernels and functions with Float, see Generic Type Annotations.