[MPlayer-users] Excessive * audio codec cpu usage

[daniel carter]

Brian J. Murrell wrote:

>   The only
>difference between:
>
>A:  17.8 V:  17.8 A-V:  0.020 ct:  0.018  528/528  58%  0% 28.0% 20 0 0%
>
>and
>A:  14.7 V:  14.8 A-V: -0.080 ct:  0.004  436/436  65%  0%  2.4% 0 0 0%
>
>is that the first one is reading the stream being written, the second
>one is reading a different stream while the other one continue to
>write.
>
Just wild speculation here. Could it be something to do with file system 
caching?

Either on the server or client.  On the server readahead may be getting 
invalidated because the file is being modified, or on the client it may 
not be caching it because the server reports that someone else has a 
write lock on the file.  Thus I/O throughput suffers and mplayer ends up 
spending allot of time waiting on reads?