3.102. __nv_fast_expf

Prototype:

float @__nv_fast_expf(float %x) 

Description:

Calculate the fast approximate base e exponential of the input argument x, e x .

Returns:

Returns an approximation to e x .

Note:

For accuracy information see the CUDA C++ Programming Guide, Mathematical Functions Appendix, Intrinsic Functions section.

Most input and output values around denormal range are flushed to sign preserving 0.0.

Library Availability:

Compute 2.0: Yes

Compute 3.0: Yes

Compute 3.5: Yes