#include <stdlib.h>         // for malloc(), free()
#include <stdio.h>          // for printf()
#include <stddef.h>         // for size_t, ptrdiff_t
#include <math.h>           // for nan()

void
initGeMatrix(size_t m, size_t n,
             double *A,
             ptrdiff_t incRowA, ptrdiff_t incColA)
{
    for (size_t i=0; i<m; ++i) {
        for (size_t j=0; j<n; ++j) {
            A[i*incRowA + j*incColA] = i*n + j +1;
        }
    }
}

void
printGeMatrix(size_t m, size_t n,
              const double *A,
              ptrdiff_t incRowA, ptrdiff_t incColA)
{
    for (size_t i=0; i<m; ++i) {
        for (size_t j=0; j<n; ++j) {
            printf("%9.2lf ", A[i*incRowA + j*incColA]);
        }
        printf("\n");
    }
    printf("\n");
}

void
dcopy(size_t n,
      const double *x, ptrdiff_t incX,
      double *y, ptrdiff_t incY)
{
    for (size_t i=0; i<n; ++i) {
        y[i*incY] = x[i*incX];
    }
}

void
daxpy(size_t n, double alpha,
      const double *x, ptrdiff_t incX,
      double *y, ptrdiff_t incY)
{
    for (size_t i=0; i<n; ++i) {
        y[i*incY] += alpha*x[i*incX];
    }
}

double
ddot(size_t n,
     const double *x, ptrdiff_t incX,
     const double *y, ptrdiff_t incY)
{
    double result = 0;

    for (size_t i=0; i<n; ++i) {
        result += x[i*incX]*y[i*incY];
    }

    return result;
}

void
dscal(size_t n,
      double alpha,
      double *x, ptrdiff_t incX)
{
    if (alpha==1) {
        return;
    }
    if (alpha!=0) {
        for (size_t i=0; i<n; ++i) {
            x[i*incX] *= alpha;
        }
    } else {
        for (size_t i=0; i<n; ++i) {
            x[i*incX] = 0;
        }
    }
}


#ifndef COLMAJOR
#define COLMAJOR 1
#endif

int
main()
{
    printf("COLMAJOR = %d\n", COLMAJOR);

    size_t m = 5, n = 10;

    ptrdiff_t incRowA = COLMAJOR ? 1 : n;
    ptrdiff_t incColA = COLMAJOR ? m : 1;

    double *A = malloc(m*n*sizeof(double));
    if (!A) {
        abort();
    }

    initGeMatrix(m, n, A, incRowA, incColA);

    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    printf("ddot(A(:,1), A(:,4)) = %lf\n",
           ddot(m, &A[1*incColA], incRowA, &A[4*incColA], incRowA));

    printf("compute: A(:,1) -= 0.5*A(:,4)\n");
    daxpy(m, -0.5, &A[4*incColA], incRowA, &A[1*incColA], incRowA);

    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    printf("copy: A(1,:) = A(2,:)\n");
    dcopy(n, &A[2*incRowA], incColA, &A[1*incRowA], incColA);

    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    printf("scal: A(1,:) = 1.5*A(1,:)\n");
    dscal(n, 1.5, &A[1*incRowA], incColA);

    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    //
    // We manually insert some NaNs into A
    //
    A[1*incRowA+0*incColA] = nan("");
    A[1*incRowA+3*incColA] = nan("");
    A[1*incRowA+9*incColA] = nan("");

    printf("\n\nAfter inserting some NaNs into A:\n");
    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    printf("scal: A(1,:) = 1.5*A(1,:)       <- will give wrong results\n");
    dscal(n, 1.5, &A[1*incRowA], incColA);

    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    printf("scal: A(1,:) = 0*A(1,:)\n");
    dscal(n, 0.0, &A[1*incRowA], incColA);

    printf("A =\n");
    printGeMatrix(m, n, A, incRowA, incColA);

    free(A);
}