0

I looked at the documentation for the Intel intrinsics and IPP but couldn't find it right away, can anyone advice whether there are vectorized conversion routines from float to int in intrinsics or IPP?

Pascal Cuoq
  • 79,187
  • 7
  • 161
  • 281
bitwise
  • 541
  • 6
  • 16

2 Answers2

0

I think you're looking for the AVX conversion operations: https://software.intel.com/sites/products/documentation/doclib/iss/2013/compiler/cpp-lin/GUID-F9D84C02-4F98-4931-8EE6-2E68D70D7624.htm

Simon Byrne
  • 7,694
  • 1
  • 26
  • 50
0

Also, IPP functions from signal processing domain "ippsConvert_" functions might help. For example, ippsConvert_32f32s_Sfs to convert from float to int, or ippsConvert_64f32s_Sfs to convert from double to int.

Sergey Khlystov
  • 276
  • 2
  • 5