Closed winnie022 closed 1 month ago
Hello,
I am new to dperf. Thank you for maintaining great tool.
- I would like to test 1 MPPS, 2 MPPS, etc by gradually increasing the rate. Is it possible to control pps rate on TX side?
yes. pps = cc * ( 1s / keepalive-interval)
I am currently explore flood under performance/tx-pps and seems like cc can control it. There are multiple parameters like cc, cps, and keepalive, slow_start. Even udp has slow start and keepalive features?
see more about keepalive https://dperf.org/doc/html/configuration#section5
Is tx-pps the best option to test it?
use 'cc', 'keepalive', etc to control it.
I am using dperf on TX side and testpmd on RX side. (one direction)
please use dperf at both side.
if you only test tx-pps, you should use 'flood' in client.conf, and server.conf use a diffrent 'listen port', so dperf server will drop all the recving packets from dperf client.
- If I can measure round-trip pps (bi-directional) to check packet loss, what is the best option in dperf and which one should I take a look under test directory?
dperf.org
- Are there any docs or articles to understand these parameters in details?
dperf.org
Thanks,
Hello,
I am new to dperf. Thank you for maintaining great tool.
I would like to test 1 MPPS, 2 MPPS, etc by gradually increasing the rate. Is it possible to control pps rate on TX side? I am currently explore flood under performance/tx-pps and seems like cc can control it. There are multiple parameters like cc, cps, and keepalive, slow_start. Even udp has slow start and keepalive features? Is tx-pps the best option to test it? I am using dperf on TX side and testpmd on RX side. (one direction)
If I can measure round-trip pps (bi-directional) to check packet loss, what is the best option in dperf and which one should I take a look under test directory?
Are there any docs or articles to understand these parameters in details?
Thanks,