Question of CUDA with visual studio

I have a problem, I’m tying to pass Host to device, vecinos is a * float has different values of 0.0f , but d_vecinos is difined int* d_Vecinos (global)
and each value of d_Vecinos is 0.0f after of cudaMemcpy.

cudaMalloc((void**) &d_Vecinos, tamanioVecinos*sizeof(int));


It compiles. This code is in method named inicializar() of main.cpp