Hello…
I’m beginner using Cuda.
I wanna efficiently use 2-D array. Which are very small or very large…
I Know col, row…
Solve Problem with (M * N) Array
solution <<< dimGrid, dimBlock >>> (PARAMETER);
If i configurate <<< m , n >>>
m = 1000, n = 700
above m,n value, This function works well ???
My GPGPU :
Threads per a block : 512
MAX-Dim of a block : 512 * 512 * 64
max-dim of a grid : 65535 * 65535 * 1
Becuase 1000 > 512, 700 > 512 , I don’t know that This function works well…
and. Solution function
for(i = 0 ; i < m ; i ++ )
for(j = 0 ; j < n ; j++ )
{
temp[i*m + j] = arr[i*m + j] * alpha;
result[i] += temp[i*m + j];
}
This forms are availabe ?? Not CPU… In GPU