[...] Why? I don't know exactly why, but I am sure that pgbench isn't doing anything weird. It's either libpq acting funny, or the OS.
My guess is the OS. "PQfinish" or "select" do/are systems calls that present opportunities to switch context. I think that the OS is passing time with other processes on the same host, expecially postgres backends, when it is not with the client. In order to test that, pgbench should run on a dedicated box with less threads than the number of available cores, or user time could be measured in addition to elapsed time. Also, testing with many clients per thread means that if any client is stuck all other clients incur an artificial latency: measures are intrinsically fragile.
I need to catch up with revisions done to this feature since I started instrumenting my copy more heavily. I hope I can get this ready for commit by Monday.
Ok, thanks! -- Fabien. -- Sent via pgsql-hackers mailing list (email@example.com) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers