Naive Matrix

We implement here the matrix multiplication algorithm. As all such projects, it performs badly compared to vendors library such as cublas. We use the Naive Matrix Class to materialize a matrix.

The implementation is naive, and no optimization effort is done:

  [EntryPoint]
  public static void ComputeRowsOfProduct(NaiveMatrix resultMatrix, NaiveMatrix matrixA, NaiveMatrix matrixB, int lineFrom, int lineTo)
  {
      int commonSize = matrixA.Width;
      int bWidth = matrixB.Width;
      for (int i = lineFrom + threadIdx.y + blockIdx.y * blockDim.y; i < lineTo; i += blockDim.y * gridDim.y)
      {
          for (int j = threadIdx.x + blockIdx.x * blockDim.x; j < bWidth; j += blockDim.x * gridDim.x)
          {
              resultMatrix[i * bWidth + j] = 0.0f;
                    
              for (int k = 0; k < commonSize; ++k)
              {
                  resultMatrix[i * bWidth + j] += (matrixA[i * commonSize + k] * matrixB[k * bWidth + j]);
              }
          }
      }
  }

Purpose of this example is to serve as base example for Shared-Matrix example, which demonstrates usage of shared memory from C#.

Newton Fractal

Home

Shared Matrix