t
Copyright (c) 2012, Oracle and/or its affiliates. All rights reserved.
Modified for Solaris to add the Solaris stability classification,
and to add a note about source availability.

ib_read_bw, ib_read_lat - Bandwith and latency test using RDMA read transactions.

ib_read_bw [-p, --port=<port>] [-d, --ib-dev=<dev>]
 [-i, --ib-port=<port number.] [-c, --connection=<RC|UD>]
 [-m, --mtu=<mtu>] [-s, --size=<size>] [-a, --all]
 [-n, --iters=<iterations>] [-t, --tx-depth=<depth>]
 [-u, --qp-timeout=<timeout] [-S, --sl=<sl>]
 [-x, --gid-index=<index>] [-F, --CPU-freq] [-V, --version]
 [-b, --bidirectional] [-e, --events] [-N, --no peak-bw]
 [-o, --outs=<num>] [server]

ib_read_lat [-p, --port=<port>] [-d, --ib-dev=<dev>]
 [-i, --ib-port=<port number.] [-c, --connection=<RC|UD>]
 [-m, --mtu=<mtu>] [-s, --size=<size>] [-a, --all]
 [-n, --iters=<iterations>] [-t, --tx-depth=<depth>]
 [-u, --qp-timeout=<timeout] [-S, --sl=<sl>]
 [-x, --gid-index=<index>] [-F, --CPU-freq] [-V, --version]
 [-e, --events] [-o, --outs=<num>] [-C, --report-cycles]
 [-H, --report-histogram] [-U, --report-unsorted] [server]

.Nm The ib_read_bw and ib_read_lat tests are part of the "perftest" utilities that are used for InfiniBand related performance testing. The "perftest" utilities are micro bench mark utilities.

First a passive receiving instance is started.

Then an active sending instance is started, giving it the address at which it will find a listening passive receiver.

If options are used then the same options must be used for both client and server.

The following options are available for use on the command line:

7 -p, --port <port> Listen on/connect to port <port> (default 18515)

-d, --ib-dev <dev> Use IB device <dev> (default first device found)

-i, --ib-port <port number> Use port <port number> of IB device (default 1)

-c, --connection <RC|UD> Connection type RC or UD (default RC)

-m, --mtu <mtu>s Mtu size : 256 - 4096 (default port mtu)

-s, --size <size> Size of message to exchange (default 2)

-a, --all Run sizes from 2 till 2^23

-n, --iters <iterations> Number of exchanges (at least 5, default 1000)

-t, --tx-depth <depth> Size of tx queue (default 50)

-u, --qp-timeout <timeout> QP timeout, timeout value is 4 usec * 2 ^(timeout), default 14

-S, --sl <sl> SL (default 0)

-x, --gid-index <index> Test uses GID with GID index (Default : IB - no gid)

-F, --CPU-freq Do not fail even if cpufreq_ondemand module is loaded

-V, --version Display version number

-b, --bidirectional Measure bidirectional bandwidth (default unidirectional)

-e, --events Sleep on CQ events (default poll)

-N, --no peak-bw Cancel peak-bw calculation (default with peak)

-o, --outs <num> Specify <num> as number of outstanding read/atom(default max of device)

-C, --report-cycles report times in cpu cycle units (default microseconds)

-H, --report-histogram Print out all results (default print summary only)

-U, --report-unsorted (implies -H) print out unsorted results (default sorted)

ib_write_bw(1), ib_send_bw(1), rdma_lat(1), ib_write_bw_postlist(1), rdma_bw(1), ib_write_lat(1), ib_send_lat(1), ib_clock_test(1)