Re: [ecasound] now that's reliable!

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [ecasound] now that's reliable!
From: S. Massy (theanaloguekid_AT_tak.net.dhis.org)
Date: Wed Oct 10 2001 - 17:40:57 EEST

On Tue, 09 Oct 2001, Kai Vehmanen <k_AT_eca.cx> wrote:

>
> Ecasound command-line (runs as root):
>
> $ ecasound_debug -i smile.mp3 -o alsa -ea:100 -c
>
> Top shows:
>
> --cut--
> 6:11pm up 1 day, 9:55, 9 users, load average: 4.11, 2.35, 1.56
> 97 processes: 88 sleeping, 8 running, 1 zombie, 0 stopped
> CPU states: 49.5% user, 50.4% system, 0.0% nice, 0.0% idle
> Mem: 384848K av, 366420K used, 18428K free, 0K shrd, 4356K buff
> Swap: 136544K av, 6812K used, 129732K free 157536K cached
>
> PID USER PRI NI SIZE RSS SHARE FLAGS STAT %CPU %MEM COMMAND
> 18533 root 20 0 8032 8024 4660 100040 R 99.9 2.0 ecasound_debug
> 1401 kaiv 19 0 36584 34M 3908 0 R 24.9 9.2 netscape-naviga
> 18532 root 12 0 8032 8024 4660 140 S 10.4 2.0 ecasound_debug
> 19035 kaiv 18 0 3276 3276 1152 0 R 6.3 0.8 cc1
> --cut--
>
> Because of the '-ea:100', ecasound chooses the 'rtlowlatency' mode, which
> means that minimal buffering is used. But even with small buffers,
> runtime performance is rock solid. And as you can see, machine is
> _heavily_ loaded (mozilla, compiling alsa-driver, compiling ecasound and
> ecasound itself). User-interface responds slowly, but that's ok as long as
> audio is ok.
Actually, I wish I could say the same. :(
There's a great difference of performance for me when it switches to
rtlowlatency, it's fine so long as the system is not overloaded, but
nointbuf makes ecasound very prone to xruns. SCHED_FIFO partly saves
the game but I definitely see a difference between rt and rtlowlatency
for simple playback/recording. On the other hand, I rarely run
ecasound as root for simple playback so it would rather default to rt
than to rtlowlatency.

>
> PS I'm using a non-patched 2.4.10 smp-kernel and ALSA-CVS ~0.9beta8.
Same here, but not SMP.
> PPS When compiled with (--enable-debug, produces ecasound_debug),
> ecasound runs _extremely_ slow (all debug-checks on, no optimizations,
> nothing inlined).
>
> --
> http://www.eca.cx
> Audio software for Linux!
>
> --
> To unsubscribe send message 'unsubscribe' in the body of the
> message to <ecasound-list-request_AT_wakkanet.fi>.
> 

--
To unsubscribe send message 'unsubscribe' in the body of the
message to <ecasound-list-request_AT_wakkanet.fi>.

New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Wed Oct 10 2001 - 17:36:21 EEST