public static double[] Mark3(int[,] A, int[,] B, int[,] C, int Size, int Size1d, int n, int count)
{
double[] result = new double[n];
double dummy = 0.0;
CudafyModule km = CudafyTranslator.Cudafy();
GPGPU gpu = CudafyHost.GetDevice(CudafyModes.Target, CudafyModes.DeviceId);
gpu.LoadModule(km);
GPGPUProperties GPU_prop = gpu.GetDeviceProperties();
int maxTheadBlockSize = (int)Math.Sqrt(GPU_prop.MaxThreadsPerBlock);
for (int j = 0; j < n; j++)
{
Timer t = new Timer();
for (int i = 0; i < count; i++)
{
dummy += MA(A, B, C, gpu, maxTheadBlockSize,Size);
}
double time = t.Check() / count;
result[j] = time;
Console.WriteLine("time: " + time + " ms");
}
return result;
}
public static double[] Mark4(int[,] A, int[,] B, int[,] C, int Size, int Size1d, int n, int count)
{
double dummy = 0.0;
double st = 0.0, sst = 0.0;
CudafyModule km = CudafyTranslator.Cudafy();
GPGPU gpu = CudafyHost.GetDevice(CudafyModes.Target, CudafyModes.DeviceId);
gpu.LoadModule(km);
GPGPUProperties GPU_prop = gpu.GetDeviceProperties();
int maxTheadBlockSize = (int)Math.Sqrt(GPU_prop.MaxThreadsPerBlock);
for (int j = 0; j < n; j++)
{
Timer t = new Timer();
for (int i = 0; i < count; i++)
dummy += MA(A, B, C, gpu, maxTheadBlockSize, Size);
double time = t.Check() / count;
st += time;
sst += time * time;
}
double mean = st / n, sdev = Math.Sqrt((sst - mean * mean * n) / (n - 1));
return new double[2] { mean, sdev };
}