| 1 |
This directory contains a few example programs. |
| 2 |
|
| 3 |
Each program takes the filename as a command-line argument |
| 4 |
"-fname filename". |
| 5 |
|
| 6 |
If you are using "mpirun" to run an MPI program, you can run the |
| 7 |
program "simple" with two processes as follows: |
| 8 |
mpirun -np 2 simple -fname test |
| 9 |
|
| 10 |
|
| 11 |
simple.c: Each process creates its own file, writes to it, reads it |
| 12 |
back, and checks the data read. |
| 13 |
|
| 14 |
psimple.c: Same as simple.c but uses the PMPI versions of all MPI routines |
| 15 |
|
| 16 |
error.c: Tests if error messages are printed correctly |
| 17 |
|
| 18 |
status.c: Tests if the status object is filled correctly by I/O functions |
| 19 |
|
| 20 |
perf.c: A simple read and write performance test. Each process writes |
| 21 |
4Mbytes to a file at a location determined by its rank and |
| 22 |
reads it back. For a different access size, change the value |
| 23 |
of SIZE in the code. The bandwidth is reported for two cases: |
| 24 |
(1) without including MPI_File_sync and (2) including |
| 25 |
MPI_File_sync. |
| 26 |
|
| 27 |
async.c: This program is the same as simple.c, except that it uses |
| 28 |
asynchronous I/O. |
| 29 |
|
| 30 |
coll_test.c: This program tests the use of collective I/O. It writes |
| 31 |
a 3D block-distributed array to a file corresponding to the |
| 32 |
global array in row-major (C) order, reads it back, and checks |
| 33 |
that the data read is correct. The global array size has been |
| 34 |
set to 32^3. If you are running it on NFS, which is very slow, |
| 35 |
you may want to reduce that size to 16^3. |
| 36 |
|
| 37 |
coll_perf.c: Measures the I/O bandwidth for writing/reading a 3D |
| 38 |
block-distributed array to a file corresponding to the global array |
| 39 |
in row-major (C) order. The global array size has been |
| 40 |
set to 128^3. If you are running it on NFS, which is very slow, |
| 41 |
you may want to reduce that size to 16^3. |
| 42 |
|
| 43 |
misc.c: Tests various miscellaneous MPI-IO functions |
| 44 |
|
| 45 |
atomicity.c: Tests whether atomicity semantics are satisfied for |
| 46 |
overlapping accesses in atomic mode. The probability of detecting |
| 47 |
errors is higher if you run it on 8 or more processes. |
| 48 |
|
| 49 |
large_file.c: Tests access to large files. Writes a 4-Gbyte file and |
| 50 |
reads it back. Run it only on one process and on a file system |
| 51 |
on which ROMIO supports large files. |
| 52 |
|
| 53 |
large_array.c: Tests writing and reading a 4-Gbyte distributed array using |
| 54 |
the distributed array datatype constructor. Works only on file |
| 55 |
systems that support 64-bit file sizes and MPI implementations |
| 56 |
that support 64-bit MPI_Aint. |
| 57 |
|
| 58 |
file_info.c: Tests the setting and retrieval of hints via |
| 59 |
MPI_File_set_info and MPI_File_get_info |
| 60 |
|
| 61 |
excl.c: Tests MPI_File_open with MPI_MODE_EXCL |
| 62 |
|
| 63 |
noncontig.c: Tests noncontiguous accesses in memory and file using |
| 64 |
independent I/O. Run it on two processes only. |
| 65 |
|
| 66 |
noncontig_coll.c: Same as noncontig.c, but uses collective I/O |
| 67 |
|
| 68 |
noncontig_coll2.c: Same as noncontig_coll.c, but exercises the |
| 69 |
cb_config_list hint and aggregation handling more. |
| 70 |
|
| 71 |
i_noncontig.c: Same as noncontig.c, but uses nonblocking I/O |
| 72 |
|
| 73 |
shared_fp.c: Tests the shared file pointer functions |
| 74 |
|
| 75 |
split_coll.c: Tests the split collective I/O functions |
| 76 |
|
| 77 |
fperf.f: Fortran version of perf.c |
| 78 |
|
| 79 |
fcoll_test.f: Fortran version of coll_test.c |
| 80 |
|
| 81 |
pfcoll_test.f: Same as fcoll_test.f but uses the PMPI versions of |
| 82 |
all MPI routines |
| 83 |
|
| 84 |
fmisc.f: Fortran version of misc.c |