Re: Sun vs. Mac - gprof output
| От | Tom Lane |
|---|---|
| Тема | Re: Sun vs. Mac - gprof output |
| Дата | |
| Msg-id | 12202.1042697508@sss.pgh.pa.us обсуждение |
| Ответ на | Re: Sun vs. Mac - gprof output (CaptainX0r <captainx0r@yahoo.com>) |
| Ответы |
Re: Sun vs. Mac - gprof output
|
| Список | pgsql-performance |
CaptainX0r <captainx0r@yahoo.com> writes:
> % cumulative self self total
> time seconds seconds calls ms/call ms/call name
> 15.8 111.61 111.61 rl_signal_handler [5]
> 4.3 142.18 30.57 rl_sigwinch_handler [6]
> 1.9 155.42 13.24 rl_set_sighandler [8]
> 1.9 168.52 13.10 rl_maybe_set_sighandler
> [9]
> 1.1 176.37 7.85 _rl_next_macro_key [11]
> 0.9 182.38 6.01 rl_read_key [12]
> 0.8 188.07 5.69 rl_backward_kill_line [13]
All of these names correspond to internal routines in libreadline.
It'd not be surprising for libreadline to suck a good deal of the
runtime of psql ... but I don't believe the backend will call it at all.
So, either this trace is erroneous, or you profiled the wrong process
(client instead of backend), or there's something truly weird going on.
regards, tom lane
В списке pgsql-performance по дате отправления: