I’m trying to use cudaMallocPitch():
r = cudaMallocPitch(udev, pitch, nx, ny*nz)
(where udev is an array of floats, and r, pitch, nx, ny, and nz are integers)
When I try to compile, I get the following error:
“PGF90-S-0155-Could not resolve generic procedure cudamallocpitch”
Apparently TheMatt got cudaMallocPitch() to work (see post from 2009: https://forums.developer.nvidia.com/t/cuda-fortran-and-cuda-api-3d-arrays/131782/1 ) and as far as I can tell, I’m using the function the same way he did. Any ideas?
I’m using pgfortran 10.2-0, should that be a problem?