Question about cudaGetErrorString return values

Accelerated Computing CUDA CUDA Programming and Performance
2 
$ cat cges.cpp
#include <cuda_runtime.h>
#include <iostream>

int main(){

  for (int i = 0; i < 999; i++)
    std::cout << i << " : " << cudaGetErrorString((cudaError_t)i) << std::endl;
}
$ g++ cges.cpp -o cges -I/usr/local/cuda/include -L/usr/local/cuda/lib64 -lcudart
$ ./cges | grep "driver shutting"
4 : driver shutting down
$

The errors are all enumerated in /usr/local/cuda/include/driver_types.h

The description for that error in driver_types.h is:

    /**
     * This indicates that a CUDA Runtime API call cannot be executed because
     * it is being called during process shut down, at a point in time after
     * CUDA driver has been unloaded.
     */
    cudaErrorCudartUnloading              =     4,
1 Like