115 lines
No EOL
2.4 KiB
C++
115 lines
No EOL
2.4 KiB
C++
#include <iostream>
|
|
#include <vector>
|
|
#include <cmath>
|
|
using namespace std;
|
|
|
|
|
|
// Inner product
|
|
double benchmark_A(const vector<double> &x, const vector<double> &y)
|
|
{
|
|
double sum = 0.0;
|
|
#pragma omp parallel for reduction(+:sum)
|
|
for (unsigned int i = 0; i < x.size(); i++)
|
|
{
|
|
sum += x[i]*y[i];
|
|
}
|
|
return sum;
|
|
}
|
|
|
|
// Inner product
|
|
double benchmark_A_sum(const vector<double> &x)
|
|
{
|
|
double sum = 0.0;
|
|
#pragma omp parallel for reduction(+:sum)
|
|
for (unsigned int i = 0; i < x.size(); i++)
|
|
{
|
|
sum += x[i];
|
|
}
|
|
return sum;
|
|
}
|
|
|
|
//Matrix-vector product
|
|
vector<double> benchmark_B(const vector<double> &A, const vector<double> &x)
|
|
{
|
|
unsigned int N = x.size();
|
|
unsigned int M = A.size() / N;
|
|
vector<double> b(M, 0.0);
|
|
|
|
#pragma omp parallel for
|
|
for (unsigned int i = 0; i < M; i++)
|
|
{
|
|
double bi = 0.0;
|
|
for (unsigned int j = 0; j < N; j++)
|
|
{
|
|
bi += A[i*N+j]*x[j];
|
|
}
|
|
b[i] = bi;
|
|
}
|
|
|
|
return b;
|
|
}
|
|
|
|
|
|
//Matrix-Matrix product
|
|
vector<double> benchmark_C(const vector<double> &A, const vector<double> &B, unsigned int M)
|
|
{
|
|
unsigned int L = A.size()/M;
|
|
unsigned int N = B.size()/L;
|
|
vector<double> C(M*N,0.0);
|
|
#pragma omp parallel for collapse(2)
|
|
for (unsigned int i = 0; i < M; i++)
|
|
{
|
|
for (unsigned int j = 0; j < N; j++)
|
|
{
|
|
double sum = 0.0;
|
|
for (unsigned int k = 0; k < L; k++)
|
|
{
|
|
sum += A[i*L+k]*B[k*N+j];
|
|
}
|
|
C[i*N+j] = sum;
|
|
}
|
|
}
|
|
|
|
return C;
|
|
|
|
}
|
|
|
|
//polynomial evaluation
|
|
vector<double> benchmark_D(const vector<double>& coeff, const vector<double>& x)
|
|
{
|
|
unsigned int p = coeff.size(); // p coefficients, degree p-1
|
|
unsigned int N = x.size();
|
|
vector<double> y(N);
|
|
|
|
#pragma omp parallel for
|
|
for (unsigned int i = 0; i < N; i++){
|
|
double yi = coeff[p-1];
|
|
double xi = x[i];
|
|
for(int j=p-2; j>=0; --j)
|
|
{
|
|
yi = yi*xi+coeff[j];
|
|
}
|
|
y[i] = yi;
|
|
}
|
|
return y;
|
|
}
|
|
|
|
|
|
double benchmark_A_old(const vector<double> &x, const vector<double> &y)
|
|
{
|
|
double sum = 0.0;
|
|
for (unsigned int i = 0; i < x.size(); i++)
|
|
{
|
|
sum += x[i]*y[i];
|
|
}
|
|
return sum;
|
|
}
|
|
double benchmark_A_sum_old(const vector<double> &x)
|
|
{
|
|
double sum = 0.0;
|
|
for (unsigned int i = 0; i < x.size(); i++)
|
|
{
|
|
sum += x[i];
|
|
}
|
|
return sum;
|
|
} |