Open TheNumbat opened 2 weeks ago
Updates the following builtins to operate on 32-bit floats:
caml_float32x4_low_to_float
caml_float32x4_low_of_float
caml_float32x4_const1
caml_float32x4_const4
The new builtins take and return unboxed float32s. This allows us to remove the special cases in the backend for converting to/from 64-bit floats.
float32
The SIMD tests have been updated to reflect the new behavior.
Updates the following builtins to operate on 32-bit floats:
caml_float32x4_low_to_float
caml_float32x4_low_of_float
caml_float32x4_const1
caml_float32x4_const4
The new builtins take and return unboxed
float32
s. This allows us to remove the special cases in the backend for converting to/from 64-bit floats.The SIMD tests have been updated to reflect the new behavior.