CUDA fmod - 不允许从__global__函数调用__host__函数

Question

I'm compiling a CUDA 5 project with VS2012 and I'm receiving this error:

error : calling a __host__ function("fmod<float, int> ") from a __global__ function is not allowed

The compiler is practically mismatching the CUDA __device__ fmod with the fmod defined in math.h

How can I "disambiguate" between them?

Answer 1

This is happening because of a template argument mismatch. You are calling fmod with a floating point argument and an integer argument. There is only a device template argument for a pair of single precision or double precision floating point arguments (see here ).

What you are probably looking for is fmod(float, (float)int) ....