Prototype:

float @__nv_fmaxf(float %, float %) 

Description:

Library Availability:

Compute 2.0: Yes

Compute 3.0: Yes

Compute 3.5: Yes