Does NPP Resize method supports 16 bit data?

I would like to know whether NPP has a plan to support Resize function that works with 16 bit data. In the documentation for NPP it mentions support for 8 bit data only. But for other geometric transformation APIS NPP is supporting 16 bit data. I’m using CUDA Toolkit V4.1 and i’m working with porting of my project from usage of Intel’s IPP library to NVIDIA’s NPP. Since we work frequently with 16 bit data only it would be better if NVIDIA could support a 16 bit version of the Resize method.


Hello Varun,

thanks for using NPP. Over the last couple of releases we have been improving NPP’s support of a complete set of data-types and color-channel configurations. Sadly for the upcoming 5.0 release we were not able to do this for all the Geometric Transform primitives, including Resize.

We’re currently planning the next release and completing the Geometric Transform primitives is one of our top priorities.


thanks for the update Frank… I really appreciate your quick response… :)