6-3
Cisco SFS InfiniBand Host Drivers User Guide for Linux
OL-12309-01
Chapter 6 uDAPL
Step 3 View the Throughput test results from the server.
The following example shows the Throughput test results:
Created an EP with ep_handle = 0x2a95f8a300
queried max_recv_dtos = 256
queried max_request_dtos = 1024
Accept issued...
Received an event on ep_handle = 0x2a95f8a300
Context = 29a
Connected!
received rmr_context = 1b3b78 target_address = 95e3a000 segment_length = 40000
Sent 7759.462 Mb in 1.0 seconds throughput = 7741.811 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.583 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.499 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.753 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.885 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.800 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.769 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.769 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.707 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7741.703 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.260 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.283 Mb/sec
Sent 7759.462 Mb in 1.0 seconds throughput = 7742.483 Mb/sec
total secs 13 throughput 7742 Mb/sec
Received an event on ep_handle = 0x2a95f8a300
Context = 29a
The notable performance result in the example is Throughput as 7.7 gigabits per second.
uDAPL Latency Test Performance
The uDAPL Latency test measures half of the round-trip latency for uDAPL sends. To perform a uDAPL
Latency test performance, perform the following steps:
Step 1 Start the Latency test on the server host. The syntax for the server is as follows:
/usr/local/topspin/bin/lat_server.x device_name iterations msg_size 0:poll/1:event
The following example shows how to start the Latency test on the server host:
host1$ /usr/local/topspin/bin/lat_server.x ib0 200000 1 0
latency server started on ib0
• ib0 is the name of the device.
• 200000 is the number of RDMAs to perform for the test.
• 1 is the size in bytes of the RDMA WRITE.
• 0 is a flag specifying whether polling or event should be used. 0 signifies polling, and 1 signifies
events.
Step 2 Start the Latency test on the client.
The syntax for the client is as follows:
/usr/local/topspin/bin/lat_client.x device_name server_name iterations msg_size
0:poll/1:event
The following example shows how to start the Latency test on the client: