3.267. __nv_scalbn

Prototype: 

double @__nv_scalbn(double %x, i32 %y)

Description:

Scale x by 2 n by efficient manipulation of the floating-point exponent.

Returns:

Returns x * 2 n .

  • __nv_scalbn( ± 0 , n) returns ± 0 .

  • __nv_scalbn(x, 0) returns x.

  • __nv_scalbn( ± ∞ , n) returns ± ∞ .

Library Availability:

Compute 2.0: Yes

Compute 3.0: Yes

Compute 3.5: Yes