Interface to the message passing library MPI.
More...
|
| subroutine, public | mpi_perf_test (comm, npow, output_unit) |
| | Tests the MPI library.
|
| |
Interface to the message passing library MPI.
- History
- JGH (02-Jan-2001): New error handling Performance tools JGH (14-Jan-2001): New routines mp_comm_compare, mp_cart_coords, mp_rank_compare, mp_alltoall JGH (06-Feb-2001): New routines mp_comm_free JGH (22-Mar-2001): New routines mp_comm_dup fawzi (04-NOV-2004): storable performance info (for f77 interface) Wrapper routine for mpi_gatherv added (22.12.2005,MK) JGH (13-Feb-2006): Flexible precision JGH (15-Feb-2006): single precision mp_alltoall
- Author
- JGH
◆ mpi_perf_test()
| subroutine, public mp_perf_test::mpi_perf_test |
( |
class(mp_comm_type), intent(in) |
comm, |
|
|
integer, intent(in) |
npow, |
|
|
integer, intent(in) |
output_unit |
|
) |
| |
Tests the MPI library.
- Parameters
-
| comm | the relevant, initialized communicator |
| npow | number of sizes to test, 10**1 .. 10**npow |
| output_unit | where to direct output |
- History
- JGH 6-Feb-2001 : Test and performance code
- Author
- JGH 1-JAN-2001
- Note
- quickly adapted benchmark code, will only work on an even number of CPUs.
Definition at line 54 of file mp_perf_test.F.