-__kernel gpu_mmul(__global double* ptrR, __global double* ptrA, __global double* ptrB, const int N, const int W)
+__kernel gpu_mmul(__global float* ptrR, __global float* ptrA, __global float* ptrB, const int N, const int W)
{
int row = get_global_id(0);
int col = get_global_id(1);
- double sum = 0;
+ float sum = 0;
for (int i = 0; i < N; i++)
{
sum += ptrA[row * W + i] * ptrB[i * W + col];