Closed cudaLang closed 7 years ago
http://docs.nvidia.com/pdf/NPP_Library_Signal.pdf does not include the DivCRev for double. Is it just missing and/or is there an NPP work around like a negative power function.
There are some trivial functions missing in NPP as I also find from time to time. In such a case I'd file a feature request to the NPP team. In the time being the simplest and fastest way to work around is to write a small kernel on your own that does exactly the operation you need.
The NPP extension method for DivCRev is not included for a device variable of type double.