/* 32-1. */
void xchange_halffield32() {
#ifdef _KOJAK_INST
#pragma pomp inst begin(xchangehalf32)
#endif
# ifdef MPI
shmem_barrier_all();
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ/2),
(float*)(HalfSpinor32 + 4*VOLUME),
(LX*LY*LZ*6), g_nb_t_up);
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ/2),
(LX*LY*LZ*6), g_nb_t_dn);
# if (defined PARALLELXT || defined PARALLELXYT || defined PARALLELXYZT)
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ + T*LY*LZ/2),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ),
(T*LY*LZ*6), g_nb_x_up);
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ + T*LY*LZ/2),
(T*LY*LZ*6), g_nb_x_dn);
# endif
# if (defined PARALLELXYT || defined PARALLELXYZT)
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ + T*LY*LZ + T*LX*LZ/2),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ + T*LY*LZ),
(T*LX*LZ*6), g_nb_y_up);
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ + T*LY*LZ),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ + T*LY*LZ + T*LX*LZ/2),
(T*LX*LZ*6), g_nb_y_dn);
# endif
# if (defined PARALLELXYZT)
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ + T*LY*LZ + T*LX*LZ + T*LX*LY/2),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ + T*LY*LZ + T*LX*LZ),
(T*LX*LY*6), g_nb_z_up);
shmem_float_put((float*)(HalfSpinor32 + 4*VOLUME + RAND/2 + LX*LY*LZ + T*LY*LZ + T*LX*LZ),
(float*)(HalfSpinor32 + 4*VOLUME + LX*LY*LZ + T*LY*LZ + T*LX*LZ + T*LX*LY/2),
(T*LX*LY*6), g_nb_z_dn);
# endif
shmem_barrier_all();
# endif /* MPI */
return;
#ifdef _KOJAK_INST
#pragma pomp inst end(xchangehalf32)
#endif
}
int main(int argc, char *argv[]) {
int n = 100;
// int n1=101;
start_pes(0);
int nn = (n-1) / _num_pes();
int n_local0 = 1 + _my_pe() * nn;
int n_local1 = 1 + (_my_pe()+1) * nn;
// allocate only local part + ghost zone of the arrays x,y
float *x, *y;
x = (float*) malloc((n_local1 - n_local0 + 2)*sizeof(float));
y = (float*) malloc((n_local1 - n_local0 + 2)*sizeof(float)); // forgot shmalloc
shmem_barrier_all();
//... // fill x, y
// fill ghost zone
if (_my_pe() > 0)
shmem_float_get(x,y,n1,1); // extra code
shmem_float_put(y,x, 1, _my_pe()-1);
shmem_barrier_all();
// do computation
float e = 0;
int i;
for (i=n_local0; i<n_local1; ++i) {
x[i] += ( y[i+1] + y[i-1] )*.5;
e += y[i] * y[i];
}
static float work[_SHMEM_REDUCE_SYNC_SIZE];
static long sync[_SHMEM_REDUCE_SYNC_SIZE];
static float el, es;
el = e;
shmem_float_sum_to_all(&es, &el, 1,
0, 0, _num_pes(), work, sync);
e = es;
// ... // output x, e
x += (n_local0 - 1);
y += (n_local0 - 1);
shfree(x);
shfree(y);
return 0;
}
int
main (int argc, char **argv)
{
int i;
int nextpe;
int me, npes;
int success1, success2, success3, success4, success5, success6, success7,
success8;
short src1[N];
int src2[N];
long src3[N];
long double src4[N];
long long src5[N];
double src6[N];
float src7[N];
char *src8;
short src9;
int src10;
long src11;
double src12;
float src13;
short *dest1;
int *dest2;
long *dest3;
long double *dest4;
long long *dest5;
double *dest6;
float *dest7;
char *dest8;
short *dest9;
int *dest10;
long *dest11;
double *dest12;
float *dest13;
shmem_init ();
me = shmem_my_pe ();
npes = shmem_n_pes ();
if (npes > 1) {
success1 = 0;
success2 = 0;
success3 = 0;
success4 = 0;
success5 = 0;
success6 = 0;
success7 = 0;
success8 = 0;
src8 = (char *) malloc (N * sizeof (char));
for (i = 0; i < N; i += 1) {
src1[i] = (short) me;
src2[i] = me;
src3[i] = (long) me;
src4[i] = (long double) me;
src5[i] = (long long) me;
src6[i] = (double) me;
src7[i] = (float) me;
src8[i] = (char) me;
}
src9 = (short) me;
src10 = me;
src11 = (long) me;
src12 = (double) me;
src13 = (float) me;
dest1 = (short *) shmem_malloc (N * sizeof (*dest1));
dest2 = (int *) shmem_malloc (N * sizeof (*dest2));
dest3 = (long *) shmem_malloc (N * sizeof (*dest3));
dest4 = (long double *) shmem_malloc (N * sizeof (*dest4));
dest5 = (long long *) shmem_malloc (N * sizeof (*dest5));
dest6 = (double *) shmem_malloc (N * sizeof (*dest6));
dest7 = (float *) shmem_malloc (N * sizeof (*dest7));
dest8 = (char *) shmem_malloc (4 * sizeof (*dest8));
dest9 = (short *) shmem_malloc (sizeof (*dest9));
dest10 = (int *) shmem_malloc (sizeof (*dest10));
dest11 = (long *) shmem_malloc (sizeof (*dest11));
dest12 = (double *) shmem_malloc (sizeof (*dest12));
dest13 = (float *) shmem_malloc (sizeof (*dest13));
for (i = 0; i < N; i += 1) {
dest1[i] = -9;
dest2[i] = -9;
dest3[i] = -9;
dest4[i] = -9;
dest5[i] = -9;
dest6[i] = -9;
dest7[i] = -9.0;
dest8[i] = -9;
}
*dest9 = -9;
*dest10 = -9;
*dest11 = -9;
*dest12 = -9;
*dest13 = -9.0;
nextpe = (me + 1) % npes;
/* Testing shmem_short_put, shmem_int_put, shmem_long_put,
shmem_longdouble_put, shmem_longlong_put, shmem_double_put,
shmem_float_put, shmem_putmem */
shmem_barrier_all ();
shmem_short_put (dest1, src1, N, nextpe);
shmem_int_put (dest2, src2, N, nextpe);
shmem_long_put (dest3, src3, N, nextpe);
shmem_longdouble_put (dest4, src4, N, nextpe);
shmem_longlong_put (dest5, src5, N, nextpe);
shmem_double_put (dest6, src6, N, nextpe);
shmem_float_put (dest7, src7, N, nextpe);
shmem_putmem (dest8, src8, N * sizeof (char), nextpe);
shmem_barrier_all ();
if (me == 0) {
for (i = 0; i < N; i += 1) {
if (dest1[i] != (npes - 1)) {
success1 = 1;
}
if (dest2[i] != (npes - 1)) {
success2 = 1;
}
if (dest3[i] != (npes - 1)) {
success3 = 1;
}
if (dest4[i] != (npes - 1)) {
success4 = 1;
}
if (dest5[i] != (npes - 1)) {
success5 = 1;
}
if (dest6[i] != (npes - 1)) {
success6 = 1;
}
if (dest7[i] != (npes - 1)) {
success7 = 1;
}
if (dest8[i] != (npes - 1)) {
success8 = 1;
}
}
if (success1 == 0)
printf ("Test shmem_short_put: Passed\n");
else
printf ("Test shmem_short_put: Failed\n");
if (success2 == 0)
printf ("Test shmem_int_put: Passed\n");
else
printf ("Test shmem_int_put: Failed\n");
if (success3 == 0)
printf ("Test shmem_long_put: Passed\n");
else
printf ("Test shmem_long_put: Failed\n");
if (success4 == 0)
printf ("Test shmem_longdouble_put: Passed\n");
else
printf ("Test shmem_longdouble_put: Failed\n");
if (success5 == 0)
printf ("Test shmem_longlong_put: Passed\n");
else
printf ("Test shmem_longlong_put: Failed\n");
if (success6 == 0)
printf ("Test shmem_double_put: Passed\n");
else
printf ("Test shmem_double_put: Failed\n");
if (success7 == 0)
printf ("Test shmem_float_put: Passed\n");
else
printf ("Test shmem_float_put: Failed\n");
if (success8 == 0)
printf ("Test shmem_putmem: Passed\n");
else
printf ("Test shmem_putmem: Failed\n");
}
shmem_barrier_all ();
/* Testing shmem_put32, shmem_put64, shmem_put128 */
if (sizeof (int) == 4) {
for (i = 0; i < N; i += 1) {
dest2[i] = -9;
dest3[i] = -9;
dest4[i] = -9;
}
success2 = 0;
success3 = 0;
success4 = 0;
shmem_barrier_all ();
shmem_put32 (dest2, src2, N, nextpe);
shmem_put64 (dest3, src3, N, nextpe);
shmem_put128 (dest4, src4, N, nextpe);
shmem_barrier_all ();
if (me == 0) {
for (i = 0; i < N; i += 1) {
if (dest2[i] != (npes - 1)) {
success2 = 1;
}
if (dest3[i] != (npes - 1)) {
success3 = 1;
}
if (dest4[i] != (npes - 1)) {
success4 = 1;
}
}
if (success2 == 0)
printf ("Test shmem_put32: Passed\n");
else
printf ("Test shmem_put32: Failed\n");
if (success3 == 0)
printf ("Test shmem_put64: Passed\n");
else
printf ("Test shmem_put64: Failed\n");
if (success4 == 0)
printf ("Test shmem_put128: Passed\n");
else
printf ("Test shmem_put128: Failed\n");
}
}
else if (sizeof (int) == 8) {
for (i = 0; i < N; i += 1) {
dest1[i] = -9;
dest2[i] = -9;
dest3[i] = -9;
}
success1 = 0;
success2 = 0;
success3 = 0;
shmem_barrier_all ();
shmem_put32 (dest1, src1, N, nextpe);
shmem_put64 (dest2, src2, N, nextpe);
shmem_put128 (dest3, src3, N, nextpe);
shmem_barrier_all ();
if (me == 0) {
for (i = 0; i < N; i += 1) {
if (dest1[i] != (npes - 1)) {
success1 = 1;
}
if (dest2[i] != (npes - 1)) {
success2 = 1;
}
if (dest3[i] != (npes - 1)) {
success3 = 1;
}
}
if (success1 == 0)
printf ("Test shmem_put32: Passed\n");
else
printf ("Test shmem_put32: Failed\n");
if (success2 == 0)
printf ("Test shmem_put64: Passed\n");
else
printf ("Test shmem_put64: Failed\n");
if (success3 == 0)
printf ("Test shmem_put128: Passed\n");
else
printf ("Test shmem_put128: Failed\n");
}
}
/* Testing shmem_iput32, shmem_iput64, shmem_iput128 */
shmem_barrier_all ();
if (sizeof (int) == 4) {
for (i = 0; i < N; i += 1) {
dest2[i] = -9;
dest3[i] = -9;
dest4[i] = -9;
}
success2 = 0;
success3 = 0;
success4 = 0;
shmem_barrier_all ();
shmem_iput32 (dest2, src2, 1, 2, N, nextpe);
shmem_iput64 (dest3, src3, 1, 2, N, nextpe);
shmem_iput128 (dest4, src4, 1, 2, N, nextpe);
shmem_barrier_all ();
if (me == 0) {
for (i = 0; i < N / 2; i += 1) {
if (dest2[i] != (npes - 1)) {
success2 = 1;
}
if (dest3[i] != (npes - 1)) {
success3 = 1;
}
if (dest4[i] != (npes - 1)) {
success4 = 1;
}
}
if (success2 == 0)
printf ("Test shmem_iput32: Passed\n");
else
printf ("Test shmem_iput32: Failed\n");
if (success3 == 0)
printf ("Test shmem_iput64: Passed\n");
else
printf ("Test shmem_iput64: Failed\n");
if (success4 == 0)
printf ("Test shmem_iput128: Passed\n");
else
printf ("Test shmem_iput128: Failed\n");
}
}
else if (sizeof (int) == 8) {
for (i = 0; i < N; i += 1) {
dest1[i] = -9;
dest2[i] = -9;
dest3[i] = -9;
}
success1 = 0;
success2 = 0;
success3 = 0;
shmem_barrier_all ();
shmem_iput32 (dest1, src1, 1, 2, N, nextpe);
shmem_iput64 (dest2, src2, 1, 2, N, nextpe);
shmem_iput128 (dest3, src3, 1, 2, N, nextpe);
shmem_barrier_all ();
if (me == 0) {
for (i = 0; i < N / 2; i += 1) {
if (dest1[i] != (npes - 1)) {
success1 = 1;
}
if (dest2[i] != (npes - 1)) {
success2 = 1;
}
if (dest3[i] != (npes - 1)) {
success3 = 1;
}
}
if (success1 == 0)
printf ("Test shmem_iput32: Passed\n");
else
printf ("Test shmem_iput32: Failed\n");
if (success2 == 0)
printf ("Test shmem_iput64: Passed\n");
else
printf ("Test shmem_iput64: Failed\n");
if (success3 == 0)
printf ("Test shmem_iput128: Passed\n");
else
printf ("Test shmem_iput128: Failed\n");
}
}
/* Testing shmem_short_iput, shmem_int_iput, shmem_long_iput,
shmem_double_iput, shmem_float_iput */
for (i = 0; i < N; i += 1) {
dest1[i] = -9;
dest2[i] = -9;
dest3[i] = -9;
dest6[i] = -9;
dest7[i] = -9;
}
success1 = 0;
success2 = 0;
success3 = 0;
success6 = 0;
success7 = 0;
shmem_barrier_all ();
shmem_short_iput (dest1, src1, 1, 2, N, nextpe);
shmem_int_iput (dest2, src2, 1, 2, N, nextpe);
shmem_long_iput (dest3, src3, 1, 2, N, nextpe);
shmem_double_iput (dest6, src6, 1, 2, N, nextpe);
shmem_float_iput (dest7, src7, 1, 2, N, nextpe);
shmem_barrier_all ();
if (me == 0) {
for (i = 0; i < N / 2; i += 1) {
if (dest1[i] != (npes - 1)) {
success1 = 1;
}
if (dest2[i] != (npes - 1)) {
success2 = 1;
}
if (dest3[i] != (npes - 1)) {
success3 = 1;
}
if (dest6[i] != (npes - 1)) {
success6 = 1;
}
if (dest7[i] != (npes - 1)) {
success7 = 1;
}
}
if (success1 == 0)
printf ("Test shmem_short_iput: Passed\n");
else
printf ("Test shmem_short_iput: Failed\n");
if (success2 == 0)
printf ("Test shmem_int_iput: Passed\n");
else
printf ("Test shmem_int_iput: Failed\n");
if (success3 == 0)
printf ("Test shmem_long_iput: Passed\n");
else
printf ("Test shmem_long_iput: Failed\n");
if (success6 == 0)
printf ("Test shmem_double_iput: Passed\n");
else
printf ("Test shmem_double_iput: Failed\n");
if (success7 == 0)
printf ("Test shmem_float_iput: Passed\n");
else
printf ("Test shmem_float_iput: Failed\n");
}
/* Testing shmem_double_p, shmem_float_p, shmem_int_p, shmem_long_p,
shmem_short_p */
shmem_barrier_all ();
shmem_short_p (dest9, src9, nextpe);
shmem_int_p (dest10, src10, nextpe);
shmem_long_p (dest11, src11, nextpe);
shmem_double_p (dest12, src12, nextpe);
shmem_float_p (dest13, src13, nextpe);
shmem_barrier_all ();
if (me == 0) {
if (*dest9 == (npes - 1))
printf ("Test shmem_short_p: Passed\n");
else
printf ("Test shmem_short_p: Failed\n");
if (*dest10 == (npes - 1))
printf ("Test shmem_int_p: Passed\n");
else
printf ("Test shmem_int_p: Failed\n");
if (*dest11 == (npes - 1))
printf ("Test shmem_long_p: Passed\n");
else
printf ("Test shmem_long_p: Failed\n");
if (*dest12 == (npes - 1))
printf ("Test shmem_double_p: Passed\n");
else
printf ("Test shmem_double_p: Failed\n");
if (*dest13 == (npes - 1))
printf ("Test shmem_float_p: Passed\n");
else
printf ("Test shmem_float_p: Failed\n");
}
shmem_barrier_all ();
shmem_free (dest1);
shmem_free (dest2);
shmem_free (dest3);
shmem_free (dest4);
shmem_free (dest5);
shmem_free (dest6);
shmem_free (dest7);
shmem_free (dest8);
shmem_free (dest9);
shmem_free (dest10);
shmem_free (dest11);
shmem_free (dest12);
shmem_free (dest13);
}
else {
printf ("Number of PEs must be > 1 to test shmem put, test skipped\n");
}
shmem_finalize ();
return 0;
}
void
sor (float **current_ptr, float **next_ptr)
{
int i, j, my_start, my_end, my_num_rows;
float *U_Curr_Above = (float *) shmalloc ((sizeof (float)) * ((int) floor (WIDTH / H))); /* 1d array holding values from bottom row of PE above */
float *U_Curr_Below = (float *) shmalloc ((sizeof (float)) * ((int) floor (WIDTH / H))); /* 1d array holding values from top row of PE below */
float *U_Send_Buffer = (float *) shmalloc ((sizeof (float)) * ((int) floor (WIDTH / H))); /* 1d array holding values that are currently being sent */
//float U_Curr_Above[(int)floor(WIDTH/H)]; /* 1d array holding values from bottom row of PE above */
//float U_Curr_Below[(int)floor(WIDTH/H)]; /* 1d array holding values from top row of PE below */
//float U_Send_Buffer[(int)floor(WIDTH/H)]; /* 1d array holding values that are currently being sent */
float W = 1.5;
//MPI_Request request;
//MPI_Status status;
//MPI_Comm_size(MPI_COMM_WORLD,&p);
//MPI_Comm_rank(MPI_COMM_WORLD,&my_rank);
my_start = get_start (my_rank);
my_end = get_end (my_rank);
my_num_rows = get_num_rows (my_rank);
/*
* Communicating ghost rows - only bother if p > 1
*/
if (p > 1)
{
/* send/receive bottom rows */
if (my_rank < (p - 1))
{
/* populate send buffer with bottow row */
for (i = 0; i < (int) floor (WIDTH / H); i++)
{
U_Send_Buffer[i] = current_ptr[my_num_rows - 1][i];
}
/* non blocking send */
//MPI_Isend(U_Send_Buffer,(int)floor(WIDTH/H),MPI_FLOAT,my_rank+1,0,MPI_COMM_WORLD,&request);
shmem_float_put (U_Curr_Above, U_Send_Buffer,
(int) floor (WIDTH / H), my_rank + 1);
}
//if (my_rank > ROOT) {
/* blocking receive */
//MPI_Recv(U_Curr_Above,(int)floor(WIDTH/H),MPI_FLOAT,my_rank-1,0,MPI_COMM_WORLD,&status);
//}
//MPI_Barrier(MPI_COMM_WORLD);
shmem_barrier_all ();
/* send/receive top rows */
if (my_rank > ROOT)
{
/* populate send buffer with top row */
for (i = 0; i < (int) floor (WIDTH / H); i++)
{
U_Send_Buffer[i] = current_ptr[0][i];
}
/* non blocking send */
//MPI_Isend(U_Send_Buffer,(int)floor(WIDTH/H),MPI_FLOAT,my_rank-1,0,MPI_COMM_WORLD,&request);
shmem_float_put (U_Curr_Below, U_Send_Buffer,
(int) floor (WIDTH / H), my_rank - 1);
}
//if (my_rank < (p-1)) {
/* blocking receive */
//MPI_Recv(U_Curr_Below,(int)floor(WIDTH/H),MPI_FLOAT,my_rank+1,0,MPI_COMM_WORLD,&status);
//}
//MPI_Barrier(MPI_COMM_WORLD);
shmem_barrier_all ();
}
/* solve next reds (i+j odd) */
for (j = my_start; j <= my_end; j++)
{
for (i = 0; i < (int) floor (WIDTH / H); i++)
{
if ((i + j) % 2 != 0)
{
next_ptr[j - my_start][i] =
get_val_par (U_Curr_Above, current_ptr, U_Curr_Below, my_rank,
i, j) + (W / 4) * (get_val_par (U_Curr_Above,
current_ptr,
U_Curr_Below,
my_rank, i - 1,
j) +
get_val_par (U_Curr_Above,
current_ptr,
U_Curr_Below,
my_rank, i + 1,
j) +
get_val_par (U_Curr_Above,
current_ptr,
U_Curr_Below,
my_rank, i,
j - 1) +
get_val_par (U_Curr_Above,
current_ptr,
U_Curr_Below,
my_rank, i,
j + 1) -
4 *
(get_val_par
(U_Curr_Above, current_ptr,
U_Curr_Below, my_rank, i,
j)) - (pow (H, 2) * f (i,
j)));
enforce_bc_par (next_ptr, my_rank, i, j);
}
}
}
/* solve next blacks (i+j) even .... using next reds */
for (j = my_start; j <= my_end; j++)
{
for (i = 0; i < (int) floor (WIDTH / H); i++)
{
if ((i + j) % 2 == 0)
{
next_ptr[j - my_start][i] =
get_val_par (U_Curr_Above, current_ptr, U_Curr_Below, my_rank,
i, j) + (W / 4) * (get_val_par (U_Curr_Above,
next_ptr,
U_Curr_Below,
my_rank, i - 1,
j) +
get_val_par (U_Curr_Above,
next_ptr,
U_Curr_Below,
my_rank, i + 1,
j) +
get_val_par (U_Curr_Above,
next_ptr,
U_Curr_Below,
my_rank, i,
j - 1) +
get_val_par (U_Curr_Above,
next_ptr,
U_Curr_Below,
my_rank, i,
j + 1) -
4 *
(get_val_par
(U_Curr_Above, next_ptr,
U_Curr_Below, my_rank, i,
j)) - (pow (H, 2) * f (i,
j)));
enforce_bc_par (next_ptr, my_rank, i, j);
}
}
}
shfree(U_Send_Buffer);
shfree(U_Curr_Below);
shfree(U_Curr_Above);
}