gentype
half_recip
(
| gentype x) |
gentype
native_recip
(
| gentype x) |
The half_recip
computes reciprocal.
native_recip
computes reciprocal over an implementation-defined
range. The maximum error is implementation-defined.
This built-in math function takes scalar or vector arguments. The generic type name gentype is used to indicate that the function can take float, float2, float4, float8, or float16 as the type for the arguments. For any specific use of these function, the actual type has to be the same for all arguments and the return type.
Functions with the half_
prefix are implemented with a minimum of
10-bits of accuracy i.e. an ULP value less than or equal to 8192 ulp.
Functions with the native__
prefix may map to one or more native device
instructions and will typically have better performance compared to the corresponding
functions (without the native__
prefix). The accuracy (and in some
cases the input range(s)) of these functions is implementation-defined.
The vector versions of the math functions operate component-wise. The description is per-component.
The built-in math functions are not affected by the prevailing rounding mode in the calling environment, and always return the same value as they would if called with the round to nearest even rounding mode.