int main(int argc, char *argv[])
{
double* matA = _mm_malloc(WIDTH*HEIGHT*sizeof(double), 64);
double* matB = _mm_malloc((WIDTH*HEIGHT)*sizeof(double), 64);
double* prod = _mm_malloc(WIDTH*HEIGHT*sizeof(double), 64);
double* prod_ref = _mm_malloc(WIDTH*HEIGHT*sizeof(double), 64);
int read_flag = read_matrix(TEST_FILENAME, prod_ref, matA, matB);
if (read_flag == 1)
printf("Cannot open test file\n");
else if (read_flag == 2)
printf("Error while reading data from test file");
else if (read_flag == 3)
printf("Error while closing the test file");
if (read_flag)
return 0;
uint64_t start = timestamp_us();
matmul_optimize(prod, matA, matB); /* run the optimization functions. */
uint64_t time = timestamp_us() - start;
if (compare_matrix(prod, prod_ref)) {
printf("%lu incorrect\n", time);
} else {
printf("%lu\n", time);
}
_mm_free(prod_ref);
_mm_free(prod);
_mm_free(matB);
_mm_free(matA);
return 0;
}
int main(int argc, char** argv) {
arg_size = 1024;
if(argc > 1) arg_size = atoi(argv[1]);
arg_cutoff_value = 64;
if(argc > 2) arg_cutoff_value = atoi(argv[2]);
if((arg_size & (arg_size - 1)) != 0 || (arg_size % 16) != 0) inncabs::error("Error: matrix size must be a power of 2 and a multiple of 16\n");
REAL *A = alloc_matrix(arg_size);
REAL *B = alloc_matrix(arg_size);
REAL *C = alloc_matrix(arg_size);
REAL *D = alloc_matrix(arg_size);
std::stringstream ss;
ss << "Strassen Algorithm (" << arg_size << " x " << arg_size
<< " matrix with cutoff " << arg_cutoff_value << ") ";
init_matrix(arg_size, A, arg_size);
init_matrix(arg_size, B, arg_size);
OptimizedStrassenMultiply_seq(D, A, B, arg_size, arg_size, arg_size, arg_size, 1);
inncabs::run_all(
[&](const std::launch l) {
OptimizedStrassenMultiply_par(l, C, A, B, arg_size, arg_size, arg_size, arg_size, 1);
return 1;
},
[&](int result) {
return compare_matrix(arg_size, C, arg_size, D, arg_size);
},
ss.str()
);
}
/* This function checks if a matrix is orthogonal
* returns -1 if it can't be checked
* returns 0 if it isn't orthogonal
* returns 1 if it is orthogonal */
int is_orthogonal(struct matrix m)
{
if(m.rows!=m.columns)
{
printf("This matrix isn't square\n");
return -1;
}
struct matrix c=traspose(m);
c=matrix_multiplication(m,c);
return compare_matrix(c,identity_matrix(c.rows));
}
void work(int size)
{
matrix a, b, result1, result2, result3;
// Allocate memory for matrices
allocate_matrix(&a, size);
allocate_matrix(&b, size);
allocate_matrix(&result1, size);
//allocate_matrix(&result2, size) ;
allocate_matrix(&result3, size) ;
// Initialize matrix elements
init_matrix(a);
init_matrix(b);
init_matrix_zero(result1);
// Perform sequential matrix multiplication
mm(a, b, result1);
//init_matrix_zero(result2);
//mm_fast(a, b, result2) ;
//if (compare_matrix(result1, result2)==1)
// printf("true") ;
//free_matrix(result2);
init_matrix_zero(result3);
mm_block(a, b, result3) ;
if (compare_matrix(result1, result3)==1)
printf("true") ;
// Print the result1 matrix
// print_matrix(result1);
free_matrix(a);
free_matrix(b);
free_matrix(result1);
free_matrix(result3);
}
void CTestsTest::c_test_matrix_turn()
{
int size_of_matrix=3;
int **matrix;
int i, j, k=0;
matrix = (int**)malloc(size_of_matrix*sizeof(int*));
for(i=0; i<size_of_matrix; i++)
{
matrix[i] = (int*)malloc(size_of_matrix*sizeof(int));
}
for(i=0; i<size_of_matrix; i++)
for(j=0; j<size_of_matrix; j++)
{
k++;
matrix[i][j]=k;
}
int **matrix_r;
matrix_r = (int**)malloc(size_of_matrix*sizeof(int*));
for(i=0; i<size_of_matrix; i++)
{
matrix_r[i] = (int*)malloc(size_of_matrix*sizeof(int));
}
k=0;
for(i=0; i<size_of_matrix; i++)
for(j=0; j<size_of_matrix; j++)
{
k++;
matrix_r[i][j]=k+size_of_matrix*2-4*j-2*i;
}
c_calc_matrix_turn(size_of_matrix, matrix);
QCOMPARE(compare_matrix(matrix, matrix_r, size_of_matrix), 0);
for(i=0; i<size_of_matrix; i++)
{
free(matrix[i]);
free(matrix_r[i]);
}
free(matrix);
free(matrix_r);
}
int main(int argc, char *argv[])
{
uint64_t* newimg = _mm_malloc(WIMAGE*HIMAGE*sizeof(uint64_t), 64);
uint64_t* newimg_ref = _mm_malloc(WIMAGE*HIMAGE*sizeof(uint64_t), 64);
uint16_t* filter = _mm_malloc(WFILTER*HFILTER*sizeof(uint16_t), 64);
uint16_t* image = _mm_malloc((WIMAGE*HIMAGE+2*PAD_ZERO)*sizeof(uint16_t), 64);
for (int i = 0; i < PAD_ZERO; i++) { /* PAD matrix2 with zero to ease programming the optimization functions. */
image[i] = 0;
}
image += PAD_ZERO;
int read_flag = read_matrix(TEST_FILENAME, newimg_ref, filter, image);
if (read_flag == 1)
printf("Cannot open test file\n");
else if (read_flag == 2)
printf("Error while reading data from test file");
else if (read_flag == 3)
printf("Error while closing the test file");
if (read_flag)
return 0;
uint64_t start = timestamp_us();
matconv_optimize(newimg, filter, image); /* run the optimization functions. */
uint64_t time = timestamp_us() - start;
if (compare_matrix(newimg, newimg_ref)) {
printf("%lu incorrect\n", time);
} else {
printf("%lu\n", time);
}
_mm_free(filter);
_mm_free(image-PAD_ZERO);
_mm_free(newimg);
_mm_free(newimg_ref);
return 0;
}
int main(int argc, char *argv[])
{
int workers = 1;
int dqsize = 100000;
int verify = 0;
char c;
while ((c=getopt(argc, argv, "w:q:h:c")) != -1) {
switch (c) {
case 'w':
workers = atoi(optarg);
break;
case 'q':
dqsize = atoi(optarg);
break;
case 'c':
verify = 1;
break;
case 'h':
usage(argv[0]);
break;
default:
abort();
}
}
if (optind == argc) {
usage(argv[0]);
exit(1);
}
int n = atoi(argv[optind]);
lace_init(workers, dqsize);
lace_startup(0, 0, 0);
REAL *A, *B, *C1, *C2;
if ((n & (n - 1)) != 0 || (n % 16) != 0) {
printf("%d: matrix size must be a power of 2"
" and a multiple of %d\n", n, 16);
return 1;
}
A = alloc_matrix(n);
B = alloc_matrix(n);
C1 = alloc_matrix(n);
C2 = alloc_matrix(n);
init_matrix(n, A, n);
init_matrix(n, B, n);
LACE_ME;
double t1=wctime();
CALL(OptimizedStrassenMultiply, C2, A, B, n, n, n, n);
double t2=wctime();
if (verify) {
matrixmul(n, A, n, B, n, C1, n);
verify = compare_matrix(n, C1, n, C2, n);
}
if (verify)
printf("WRONG RESULT!\n");
else {
printf("Time: %f\n", t2-t1);
}
lace_exit();
return 0;
}
int main(int argc, char *argv[]) {
int n=get_num_ind();
int i,j;
struct timeval tv1,tv2;
adouble *xad;
adouble fad;
double f;
double *x;
x=new double[n];
xad=new adouble[n];
get_initial_value(x);
printf("evaluating the function...");
trace_on(tag);
for(i=0;i<n;i++)
{
xad[i] <<= x[i];
}
fad=func_eval(xad);
fad >>= f;
trace_off();
printf("done!\n");
// printf("function value =<%10.20f>\n",f);
// function(tag,1,n,x,&f);
// printf("adolc func value=<%10.20f>\n",f);
//tape_doc(tag,1,n,x,&f);
#ifdef _compare_with_full
double **H;
H = myalloc2(n,n);
printf("computing full hessain....");
gettimeofday(&tv1,NULL);
hessian(tag,n,x,H);
printf("done\n");
gettimeofday(&tv2,NULL);
printf("Computing the full hessian cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#ifdef _PRINTOUT
for(i=0;i<n;i++){
for(j=0;j<n;j++){
printf("H[%d][%d]=<%10.10f>",i,j,H[i][j]);
}
printf("\n");
}
printf("\n");
#endif
#endif
#ifdef edge_pushing
unsigned int *rind = NULL;
unsigned int *cind = NULL;
double *values = NULL;
int nnz;
int options[2];
options[0]=PRE_ACC;
options[1]=COMPUT_GRAPH;
gettimeofday(&tv1,NULL);
// edge_hess(tag, 1, n, x, &nnz, &rind, &cind, &values, options);
sparse_hess(tag,n,0,x, &nnz, &rind, &cind, &values, options);
gettimeofday(&tv2,NULL);
printf("Sparse Hessian: edge pushing cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#ifdef _PRINTOUT
for(i=0;i<nnz;i++){
printf("<%d,%d>:<%10.10f>\n",cind[i],rind[i],values[i]);
// printf("%d %d \n", rind[i], cind[i]);
}
#endif
#endif
#ifdef _compare_with_full
#ifdef edge_pushing
compare_matrix(n,H,nnz,cind,rind,values);
#endif
myfree2(H);
#endif
#ifdef edge_pushing
printf("nnz=%d\n", nnz);
free(rind); rind=NULL;
free(cind); cind=NULL;
free(values); values=NULL;
#endif
delete[] x;
delete[] xad;
return 0;
}
int main(int argc, char *argv[]) {
int n=NUM_IND;
int i,j;
struct timeval tv1,tv2;
adouble *xad;
adouble fad;
double f;
double *x;
x=new double[n];
xad=new adouble[n];
get_initials(x, n);
// printf("evaluating the function...");
trace_on(tag);
for(i=0;i<n;i++)
{
xad[i] <<= x[i];
}
fad=eval_func<adouble>(xad, n);
fad >>= f;
trace_off();
// printf("done!\n");
std::cout << "y = " << f << std::endl;
#ifdef COMPARE_WITH_FULL_HESS
double **H;
H = myalloc2(n,n);
printf("computing full hessain....");
gettimeofday(&tv1,NULL);
hessian(tag,n,x,H);
printf("done\n");
gettimeofday(&tv2,NULL);
printf("Computing the full hessian cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#ifdef PRINT_RESULTS
for(i=0;i<n;i++){
for(j=0;j<n;j++){
printf("H[%d][%d]=<%10.10f>",i,j,H[i][j]);
}
printf("\n");
}
printf("\n");
#endif
#endif
unsigned int *rind = NULL;
unsigned int *cind = NULL;
double *values = NULL;
int nnz;
int options[2];
#ifdef LIVARH
options[0]=0;
options[1]=1;
gettimeofday(&tv1,NULL);
edge_hess(tag, 1, n, x, &nnz, &rind, &cind, &values, options);
gettimeofday(&tv2,NULL);
printf("Sparse Hessian: LivarH cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#endif
#ifdef LIVARHACC
options[0]=1;
options[1]=1;
gettimeofday(&tv1,NULL);
edge_hess(tag, 1, n, x, &nnz, &rind, &cind, &values, options);
gettimeofday(&tv2,NULL);
printf("Sparse Hessian: LivarHACC cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#endif
// Sparse ADOL-C drivers report the upper matrix
#ifdef DIRECT
options[0]=0;
options[1]=1;
gettimeofday(&tv1,NULL);
sparse_hess(tag, n, 0, x, &nnz, &cind, &rind, &values, options);
gettimeofday(&tv2,NULL);
printf("Sparse Hessian: direct recovery cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#endif
#ifdef INDIRECT
options[0]=0;
options[1]=0;
gettimeofday(&tv1,NULL);
sparse_hess(tag, n, 0, x, &nnz, &cind, &rind, &values, options);
gettimeofday(&tv2,NULL);
printf("Sparse Hessian: indirect recovery cost %10.6f seconds\n",(tv2.tv_sec-tv1.tv_sec)+(double)(tv2.tv_usec-tv1.tv_usec)/1000000);
#endif
#ifdef PRINT_RESULTS
for(i=0;i<nnz;i++){
printf("<%d,%d>:<%10.10f>\n",rind[i],cind[i],values[i]);
}
#endif
#ifdef COMPARE_WITH_FULL_HESS
compare_matrix(n,H,nnz,rind,cind,values);
myfree2(H);
#endif
free(rind); rind=NULL;
free(cind); cind=NULL;
free(values); values=NULL;
delete[] x;
delete[] xad;
return 0;
}