3.70. __nv_drcp_ru

Prototype: 

double @__nv_drcp_ru(double %x)

Description:

Compute the reciprocal of x in round-up (to positive infinity) mode.

Returns:

Returns 1 x .

Note:

For accuracy information see the CUDA C++ Programming Guide, Mathematical Functions Appendix, Double-Precision Floating-Point Functions section.

Requires compute capability >= 2.0.

Library Availability:

Compute 2.0: Yes

Compute 3.0: Yes

Compute 3.5: Yes