<div dir="ltr">Hi,<br><div class="gmail_quote"><div dir="ltr">I modified/instrumented the OSU MP latency benchmark to measure time taken by each transaction in order to get min and max latencies, in addition to the average that's reported currently.<br>
I noticed that some of the transactions, albeit few, took > 100usec and > 1msec.<br>
<br>Does anybody have any ideas about what could be causing such large round trip times (>1msec) for a few transactions while the average looks pretty good ( 10usec ranges) ?<br>Is it network or system issues?<br><br>
Here is a snapshot of the output and attached is the modified code. <br>
<br>i'm using OFED 1.3, CentOS 5 and openmpi-1.2.5.<br><br>Any insights or ideas would very helpful. Thanks in advance.<br><br>Below shows # of transactions over 100usec and 1msec.<br>Iteration count was set to 60000 for each test.<br>
Latency is round trip time.<br><br>--------------------------------------------------------------------------<br># OSU MPI Latency Test v3.0<br># Size Latency (us)<br>0 13.30 over_100usec: 13 over_1msec: 2 i 601000<br>
1 13.51 over_100usec: 12 over_1msec: 0 i 601000<br>2 13.51 over_100usec: 13 over_1msec: 1 i 601000<br>4 13.81 over_100usec: 42 over_1msec: 33 i 601000<br>
8 13.92 over_100usec: 36 over_1msec: 25 i 601000<br>16 13.90 over_100usec: 10 over_1msec: 0 i 601000<br>32 14.14 over_100usec: 54 over_1msec: 44 i 601000<br>
64 14.32 over_100usec: 11 over_1msec: 1 i 601000<br>128 15.38 over_100usec: 10 over_1msec: 0 i 601000<br>256 15.94 over_100usec: 14 over_1msec: 2 i 601000<br>
512 16.74 <b>over_100usec: 77 over_1msec: 65 </b>i 601000<br>1024 21.07 over_100usec: 17 over_1msec: 0 i 601000<br>2048 24.05 over_100usec: 17 over_1msec: 1 i 601000<br>
4096 29.99 over_100usec: 37 over_1msec: 5 i 601000<br>8192 41.71 over_100usec: 39 over_1msec: 0 i 601000<br><br><br><br></div>
</div><br></div>