Simple Driver API sample don't work

Accelerated Computing CUDA CUDA Programming and Performance
1

This is simple code don’t work

#include <stdio.h>
#include <stdlib.h>
#include <cuda.h>
#include <cutil.h>

int main() {
CUresult result;
CUdevice device;
CUcontext context;

cuInit(0);

cuDeviceGet(&device, 0);
cuCtxCreate(&context, 0, device);

unsigned int free, total;

result = cuMemGetInfo(&free, &total);

cuCtxDetach(context);

printf("GPU Memory status: %10d %10d\n", free, total);

}

1>MemoryTest.obj : error LNK2019: unresolved external symbol _cuCtxDetach@4 referenced in function _main
1>MemoryTest.obj : error LNK2019: unresolved external symbol _cuMemGetInfo@8 referenced in function _main
1>MemoryTest.obj : error LNK2019: unresolved external symbol _cuCtxCreate@12 referenced in function _main
1>MemoryTest.obj : error LNK2019: unresolved external symbol _cuDeviceGet@8 referenced in function _main
1>MemoryTest.obj : error LNK2019: unresolved external symbol _cuInit@4 referenced in function _main

May be before using driver API i need any to initialize this driver API or add any compile keys for nvcc?

P.S. Sample from SDK deviceQueryDrv is the absolutely like to my code but it work, why?

2

you need to link with -lcuda (Linux) or cuda.lib (windows)

3

Thank you. I don’t notice include library in VS Project:) Resolved!