The following command is used to test the performance:
$ time smtp-source -c -d -l 32000 -m 10000 -N -s 30 127.0.0.1:2525
This command is run two times, the first time to "warm-up", the second time to measure.
Note that "user" and "sys" times relate to the smtp-source process, and not the smtpd process.
real 0m2.693s
user 0m0.291s
sys 0m1.239s
real 0m2.763s
user 0m0.198s
sys 0m0.914s
real 0m1.998s
user 0m0.252s
sys 0m0.742s
real 0m7.095s
user 0m0.342s
sys 0m1.419s
real 0m6.114s
user 0m0.196s
sys 0m1.059s
real 0m9.006s
user 0m0.233s
sys 0m1.022s
real 0m7.282s
user 0m0.175s
sys 0m0.587s
real 0m7.197s
user 0m0.231s
sys 0m1.377s
real 0m3.370s
user 0m0.312s
sys 0m1.595s