if (*dptr == MPI_BYTE)
stop = (*len)/(2 * sizeof(double));
- else
+ else
stop = *len;
for (i = 0; i < stop; i++) {
c.real = inout->real * in->real - inout->imag * in->imag;
c.imag = inout->real * in->imag + inout->imag * in->real;
*inout = c;
- in++;
+ in++;
inout++;
}
buf0 = (int *) malloc (buf_size * nprocs * sizeof(int));
assert (buf0);
- for (i = 0; i < buf_size * nprocs; i++)
+ for (i = 0; i < buf_size * nprocs; i++)
buf0[i] = rank;
#ifdef RUN_ALLTOALLV
if (rank == 0) {
buf1 = (int *) malloc (buf_size * nprocs * sizeof(int));
assert (buf1);
- for (i = 0; i < buf_size * nprocs; i++)
+ for (i = 0; i < buf_size * nprocs; i++)
buf1[i] = i;
displs = (int *) malloc (nprocs * sizeof(int));
#endif
#ifdef RUN_BARRIER
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Barrier (comm);
#endif
#ifdef RUN_BCAST
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Bcast (buf0, buf_size, MPI_INT, 0, comm);
#endif
#ifdef RUN_GATHER
- for (i = 0; i < OP_COUNT; i++)
- MPI_Gather (&buf0[rank*buf_size], buf_size,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Gather (&buf0[rank*buf_size], buf_size,
MPI_INT, buf1, buf_size, MPI_INT, 0, comm);
#endif
#ifdef RUN_SCATTER
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scatter (buf1, buf_size, MPI_INT, buf0, buf_size, MPI_INT, 0, comm);
#endif
#ifdef RUN_GATHERV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Gatherv (&buf0[rank*buf_size],
- (rank < buf_size) ? rank : buf_size,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Gatherv (&buf0[rank*buf_size],
+ (rank < buf_size) ? rank : buf_size,
MPI_INT, buf1, rcounts, displs, MPI_INT, 0, comm);
#endif
#ifdef RUN_SCATTERV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Scatterv (buf1, counts, displs, MPI_INT, buf0,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Scatterv (buf1, counts, displs, MPI_INT, buf0,
(rank < buf_size) ? rank : buf_size, MPI_INT, 0, comm);
#endif
#ifdef RUN_REDUCE
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_MAX, 0, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_MIN, 0, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_SUM, 0, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_PROD, 0, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_LAND, 0, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_BAND, 0, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_LOR, 0, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_BOR, 0, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_LXOR, 0, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_INT, MPI_BXOR, 0, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (a, answer, buf_size, ctype, user_op, 0, comm);
#endif
#endif
#ifdef RUN_ALLGATHER
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allgather (buf0, buf_size, MPI_INT, buf1, buf_size, MPI_INT, comm);
#endif
#ifdef RUN_ALLTOALL
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Alltoall (buf1, buf_size, MPI_INT, buf0, buf_size, MPI_INT, comm);
#endif
#ifdef RUN_ALLGATHERV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allgatherv (buf0,
- (rank < buf_size) ? rank : buf_size,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allgatherv (buf0,
+ (rank < buf_size) ? rank : buf_size,
MPI_INT, buf1, rcounts, displs, MPI_INT, comm);
#endif
#ifdef RUN_ALLTOALLV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Alltoallv (buf1, alltoallvcounts, displs, MPI_INT,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Alltoallv (buf1, alltoallvcounts, displs, MPI_INT,
buf0, alltoallvcounts, displs, MPI_INT, comm);
#endif
#ifdef RUN_ALLREDUCE
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_MAX, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_MIN, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_SUM, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_PROD, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_LAND, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_BAND, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_LOR, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_BOR, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_LXOR, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_INT, MPI_BXOR, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (a, answer, buf_size, ctype, user_op, comm);
#endif
#endif
#ifdef RUN_REDUCE_SCATTER
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_MAX, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_MIN, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_SUM, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_PROD, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_LAND, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_BAND, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_LOR, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_BOR, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_LXOR, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_INT, MPI_BXOR, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (a, answer, rcounts, ctype, user_op, comm);
#endif
#endif
#ifdef RUN_SCAN
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_MAX, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_MIN, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_SUM, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_PROD, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_LAND, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_BAND, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_LOR, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_BOR, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_LXOR, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_INT, MPI_BXOR, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (a, answer, buf_size, ctype, user_op, comm);
#endif
#endif
int *ricounts, *rdcounts;
#ifdef RUN_BARRIER
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Barrier (comm);
#endif
#ifdef RUN_BCAST
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Bcast (buf0, buf_size * sizeof(int), MPI_BYTE, 0, comm);
#endif
#ifdef RUN_GATHER
- for (i = 0; i < OP_COUNT; i++)
- MPI_Gather (&buf0[rank*buf_size], buf_size * sizeof(int),
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Gather (&buf0[rank*buf_size], buf_size * sizeof(int),
MPI_BYTE, buf1, buf_size * sizeof(int), MPI_BYTE, 0, comm);
#endif
#ifdef RUN_SCATTER
- for (i = 0; i < OP_COUNT; i++)
- MPI_Scatter (buf1, buf_size * sizeof(int), MPI_BYTE,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Scatter (buf1, buf_size * sizeof(int), MPI_BYTE,
buf0, buf_size * sizeof(int), MPI_BYTE, 0, comm);
#endif
#ifdef RUN_GATHERV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Gatherv (&buf0[rank*buf_size],
- ((rank < buf_size) ? rank : buf_size) * sizeof(int),
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Gatherv (&buf0[rank*buf_size],
+ ((rank < buf_size) ? rank : buf_size) * sizeof(int),
MPI_BYTE, buf1, rcounts, displs, MPI_BYTE, 0, comm);
#endif
#ifdef RUN_SCATTERV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Scatterv (buf1, counts, displs, MPI_BYTE, buf0,
- ((rank < buf_size) ? rank : buf_size) * sizeof(int),
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Scatterv (buf1, counts, displs, MPI_BYTE, buf0,
+ ((rank < buf_size) ? rank : buf_size) * sizeof(int),
MPI_BYTE, 0, comm);
#endif
#ifdef RUN_REDUCE
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_MAX, 0, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_MIN, 0, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_SUM, 0, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_PROD, 0, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LAND, 0, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
- MPI_Reduce (buf0, buf1,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Reduce (buf0, buf1,
buf_size * sizeof(int), MPI_BYTE, MPI_BAND, 0, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LOR, 0, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
- MPI_Reduce (buf0, buf1,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Reduce (buf0, buf1,
buf_size * sizeof(int), MPI_BYTE, MPI_BOR, 0, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LXOR, 0, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
- MPI_Reduce (buf0, buf1,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Reduce (buf0, buf1,
buf_size * sizeof(int), MPI_BYTE, MPI_BXOR, 0, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
- MPI_Reduce (a, answer,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Reduce (a, answer,
buf_size * 2 * sizeof(double), MPI_BYTE, user_op, 0, comm);
#endif
#endif
buf1 = (int *) malloc (buf_size * nprocs * sizeof(int));
assert (buf1);
- for (i = 0; i < buf_size * nprocs; i++)
+ for (i = 0; i < buf_size * nprocs; i++)
buf1[i] = i;
displs = (int *) malloc (nprocs * sizeof(int));
#endif
#ifdef RUN_ALLGATHER
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allgather (buf0, buf_size * sizeof(int), MPI_BYTE,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allgather (buf0, buf_size * sizeof(int), MPI_BYTE,
buf1, buf_size * sizeof(int), MPI_BYTE, comm);
#endif
#ifdef RUN_ALLTOALL
- for (i = 0; i < OP_COUNT; i++)
- MPI_Alltoall (buf1, buf_size * sizeof(int), MPI_BYTE,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Alltoall (buf1, buf_size * sizeof(int), MPI_BYTE,
buf0, buf_size * sizeof(int), MPI_BYTE, comm);
#endif
#ifdef RUN_ALLGATHERV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allgatherv (buf0,
- ((rank < buf_size) ? rank : buf_size) * sizeof(int),
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allgatherv (buf0,
+ ((rank < buf_size) ? rank : buf_size) * sizeof(int),
MPI_BYTE, buf1, rcounts, displs, MPI_BYTE, comm);
#endif
#ifdef RUN_ALLTOALLV
- for (i = 0; i < OP_COUNT; i++)
- MPI_Alltoallv (buf1, alltoallvcounts, displs, MPI_BYTE,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Alltoallv (buf1, alltoallvcounts, displs, MPI_BYTE,
buf0, alltoallvcounts, displs, MPI_BYTE, comm);
#endif
#ifdef RUN_ALLREDUCE
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_MAX, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_MIN, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_SUM, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_PROD, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LAND, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allreduce (buf0, buf1,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allreduce (buf0, buf1,
buf_size * sizeof (int), MPI_BYTE, MPI_BAND, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LOR, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allreduce (buf0, buf1,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allreduce (buf0, buf1,
buf_size * sizeof (int), MPI_BYTE, MPI_BOR, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Allreduce (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LXOR, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allreduce (buf0, buf1,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allreduce (buf0, buf1,
buf_size * sizeof (int), MPI_BYTE, MPI_BXOR, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
- MPI_Allreduce (a, answer,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Allreduce (a, answer,
buf_size * 2 * sizeof (double), MPI_BYTE, user_op, comm);
#endif
#endif
#ifdef RUN_REDUCE_SCATTER
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_MAX, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_MIN, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_SUM, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_PROD, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_LAND, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_BYTE, MPI_BAND, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_LOR, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_BYTE, MPI_BOR, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, ricounts, MPI_UNSIGNED, MPI_LXOR, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (buf0, buf1, rcounts, MPI_BYTE, MPI_BXOR, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Reduce_scatter (a, answer, rdcounts, MPI_BYTE, user_op, comm);
#endif
#endif
#ifdef RUN_SCAN
#ifdef RUN_MAX
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_MAX, comm);
#endif
#ifdef RUN_MIN
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_MIN, comm);
#endif
#ifdef RUN_SUM
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_SUM, comm);
#endif
#ifdef RUN_PROD
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_PROD, comm);
#endif
#ifdef RUN_LAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LAND, comm);
#endif
#ifdef RUN_BAND
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size * sizeof(int), MPI_BYTE, MPI_BAND, comm);
#endif
#ifdef RUN_LOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LOR, comm);
#endif
#ifdef RUN_BOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size * sizeof(int), MPI_BYTE, MPI_BOR, comm);
#endif
#ifdef RUN_LXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size, MPI_UNSIGNED, MPI_LXOR, comm);
#endif
#ifdef RUN_BXOR
- for (i = 0; i < OP_COUNT; i++)
+ for (i = 0; i < OP_COUNT; i++)
MPI_Scan (buf0, buf1, buf_size * sizeof(int), MPI_BYTE, MPI_BXOR, comm);
#endif
#ifdef RUN_USEROP
- for (i = 0; i < OP_COUNT; i++)
- MPI_Scan (a, answer,
+ for (i = 0; i < OP_COUNT; i++)
+ MPI_Scan (a, answer,
buf_size * 2 * sizeof(double), MPI_BYTE, user_op, comm);
#endif
#endif