User guide

A–Benchmark Programs
Benchmark 3: Messaging Rate Microbenchmarks
IB0054606-02 A A-9
The benchmark has been updated with code to dynamically determine what
processes are on which host. Thefollowing is an example output when running
mpi_multibw:
$ mpirun -H host1,host2 -npernode 12 \
/usr/mpi/gcc/openmpi-1.4.3-qlc/tests/qlogic/mpi_multibw
# PathScale Modified OSU MPI Bandwidth Test
(OSU Version 2.2, PathScale $Revision: 1.1.2.1 $)
# Running on 12 procs per node (uni-directional traffic for
each process pair)
# Size Aggregate Bandwidth (MB/s) Messages/s
1 24.992623 24992622.996615
2 50.015847 25007923.312888
4 100.075479 25018869.818990
8 200.115037 25014379.610716
16 284.475601 17779725.040265
32 568.950239 17779694.953511
64 1137.899392 17779677.998115
128 1758.183987 13735812.394705
256 2116.159352 8266247.468294
512 2355.027827 4599663.724469
1024 2496.960650 2438438.134886
2048 2574.260975 1256963.366877
4096 2567.861960 626919.423819
8192 2746.514440 335267.875961
16384 3284.264487 200455.596122
32768 3292.007839 100464.106405
65536 3299.800622 50350.961641
131072 3306.998105 25230.393259
262144 3309.840069 12626.037860
524288 3323.339300 6338.766671
1048576 3323.068802 3169.125368
2097152 3307.077899 1576.937627
4194304 3300.327382 786.859365
Searching for N/2 bandwidth. Maximum Bandwidth of 3323.339300
MB/s...
Found N/2 bandwidth of 1662.009095 MB/s at size 121 bytes