culaInsufficientComputeCapability error for getri function

[golec_karolina]

My problem occurs when I am trying to invert a 9x9 culaDouble matrix. Here is the code for inversion and checking existence of errors. Unfortunately, even when I tried to print the entire error, it was blank... Why is this error thrown out and what can I do to avoid it?

Code: Select all

  M = (culaDouble*)malloc(9*9*sizeof(culaDouble));
  status = culaInitialize();
  if(status){
    culaShutdown();
    exit(EXIT_FAILURE);
  }

  memset(M, 0, 9*9*sizeof(culaDouble));

  <setting up values of M matrix>

  culaInt *ind = NULL;

  ind = (culaInt*)malloc(9*sizeof(culaInt));
  memset(ind , 0, 9*sizeof(culaInt));
   
  if(status){
    culaShutdown();
    exit(EXIT_FAILURE);
  }
   
  status = culaDgetri(9, A, 9, ind);

  if(status){
    culaShutdown();
    exit(EXIT_FAILURE);
  }



I am working in Visual Studio 2010 Professional, on Windows 7 Pro. CUDA is installed on the latest update (5.5). My GPU is Nvidia Quadro FX570, but for now, as you can see, I am not using the culaDeviceDgetri function, but the one working on the host.