how to use culaDeviceSgels in C

[sawyer]

Hi, I'm new in CULA and get some troubles.
1. the matrix is in col major,i want get the x where Ax=B in least square solution.A demension is m×n,so x is n×1 ,B is m×1. The problem is I copy A and B to device and set them in culaDeviceSgels , how I get x. The dimension of B is different from x. I dont know how to copy x to the host.

Code: Select all

   //A ,B is given,and m>=n
        float* d_A;
   float* d_B;
   cudaMalloc((void**)&d_A,sizeof(float)*m*n); //allocate device memory for d_A
   cudaMalloc((void**)&d_B,m*sizeof(float)); //I allocate device memory  for d_B
   cudaMemcpy(d_A,A,sizeof(float)*m*n,cudaMemcpyHostToDevice);//copy
   cudaMemcpy(d_B,B,sizeof(float)*m,cudaMemcpyHostToDevice);//copy
   culaInitialize();
   culaDeviceSgels('N',m,n,1,A,m,B,m);
         //Then how I copy B and get x??????
   cudaFree(d_A);
   cudaFree(d_B);
   culaShutdown();

2.if A and B are both in row-major ,how I call the function?