Garbage Value Matrix multiplication using shared memory

Accelerated Computing CUDA CUDA Programming and Performance
1

i am doing the matrix multiplication using shared memory. The algorithm is based on David krik lectures and on building the program i am not getting any errors. but in output it’s showing some garbage value. Even with my best efforts i am unable to find the error in algorithm.

please help me. :">

the kernel function is as follows.

global_ void MatrixMulKernel(int *Md, int *Nd, int *Pd, int Width)

{

device shared int Mds[2][2];

device shared int Nds[2][2];

int bx = blockIdx.x;

int by= blockIdx.y;

int tx = threadIdx.x;

int ty = threadIdx.y;

// Identify the row and column of the Pd element to work on

int Row = by * 2 + ty;

int Col = bx * 2 + tx;

int Pvalue = 0;

// Loop over the Md and Nd tiles required to compute the Pd element

 for (int m = 0; m <(Width/2); ++m) 

 {

// Coolaborative loading of Md and Nd tiles into shared memory

 Mds[ty][tx] = Md[Row*Width+(m*2+tx)];

 Nds[ty][tx] = Nd[Col+(m*2+ty)*Width];

__syncthreads();

for (int k = 0; k < 2; ++k)

 {

 Pvalue += Mds[ty][k] * Nds[k][tx];

}

 Pd[Row*Width+Col] = Pvalue;

__syncthreads();

}

}