public void Test_BLAS2_SYR2()
{
ClearBuffer(hiMatrixANN);
ClearBuffer(hiVectorXN);
ClearBuffer(hiVectorYN);
FillBuffer(hiMatrixANN);
ConverToSymmetric(hiMatrixANN, N);
FillBuffer(hiVectorXN);
FillBuffer(hiVectorYN);
diMatrixA = _gpu.Allocate(hiMatrixANN);
diVectorXN = _gpu.CopyToDevice(hiVectorXN);
diVectorYN = _gpu.CopyToDevice(hiVectorYN);
// Lower fill mode
_gpu.CopyToDevice(hiMatrixANN, diMatrixA);
_blas.SYR2(N, Alpha, diVectorXN, diVectorYN, diMatrixA);
_gpu.CopyFromDevice(diMatrixA, gpuResultNN);
for (int i = 0; i < N; i++)
{
for (int j = 0; j <= i; j++)
{
double cpuResult = Alpha * (hiVectorXN[i] * hiVectorYN[j] + hiVectorYN[i] * hiVectorXN[j]) + hiMatrixANN[GetIndexColumnMajor(i, j, N)];
Assert.AreEqual(cpuResult, gpuResultNN[GetIndexColumnMajor(i, j, N)]);
}
}
// Upper fill mode
_gpu.CopyToDevice(hiMatrixANN, diMatrixA);
_blas.SYR2(N, Alpha, diVectorXN, diVectorYN, diMatrixA, cublasFillMode.Upper);
_gpu.CopyFromDevice(diMatrixA, gpuResultNN);
for (int i = 0; i < N; i++)
{
for (int j = i; j < N; j++)
{
double cpuResult = Alpha * (hiVectorXN[i] * hiVectorYN[j] + hiVectorYN[i] * hiVectorXN[j]) + hiMatrixANN[GetIndexColumnMajor(i, j, N)];
Assert.AreEqual(cpuResult, gpuResultNN[GetIndexColumnMajor(i, j, N)]);
}
}
_gpu.FreeAll();
}