Hello Members,
I am new to cuda computing paradigm. we have an application built on eclipse IDE using c/c++ and openCV. To increase the performance of the application, we have decided to port some function calls (fft & ifft) onto the device. Anyway we decided to testto use “cufft.h” library from “/usr/local/cuda-8.0/include/” to do our fft & ifft work. As I am tried to include the library and headfile into eclipse environment and build the applicaiton, I am getting an error “make: *** [test_2dfft_v0.1] Error 1”. I created a file for the device as fft.cu and just tried to include the header as:
#include <cufft.h>
…and I am getting the following error.
Building target: test_2dfft_v0.1
Invoking: GCC C++ Linker
g++ -L/usr/local/lib/ -L/usr/share/doc/cuda-cufft-8-0 -L/usr/local/cuda-8.0/lib64/libcufft.so -L"{lib, lib64}/libcufft.so" -L/usr/local/cuda/include -L"/usr/local/cuda/lib64" -L/usr/local/cuda/bin/ -L/usr/local/cuda/lib64/ -o “test_2dfft_v0.1” ./2dfft.o ./fft.o ./main.o -lopencv_core -lcuda-cufft-8-0 -lcufft -lm -lcuda -lcudart
g++: error: ./fft.o: No such file or directory
make: *** [test_2dfft_v0.1] Error 1
Would anybody guide me into proper path?
Thanks in advance,
Venkat