AW: [MPlayer-users] rtc wasting 15% CPU

[The Wanderer]

RC wrote:

> On Sat, 7 Jan 2006 11:01:42 +0100 "Ethan Arnold" <ethan at arnold.net>
> wrote:
> 
>> Interesting, on my system there is absolutely no difference between
>> with or without -nortc. Mplayer CPU is always <1%

On my system, playing one particular file I chose randomly to test,
MPlayer takes 30%-45% of the CPU (according to gkrellm), but the
differences in the values reported by time are small enough that I think
they're negligible. (I tested both with bash's builtin time and with
/usr/bin/time, the latter being what you apparently used; with or
without -nortc I get a CPU percentage of 28%.) I can provide the actual
numbers if anyone's interested.

> The obvious question, then; is RTC setup correctly on your system?
> Does mplayer -v print a line such as:
>   "Using Linux hardware RTC timing" or
>   "Using nanosleep timing"

I get "Using Linux hardware RTC timing (1024Hz).". Adding -nortc does
indeed change that to 'Using nanosleep() timing.". It doesn't produce an
obviously significant difference in CPU usage, though.

Either there's something weird about your system, or perhaps the file
I'm using to test is sufficiently complicated that the effects of timing
issues get swamped in the rest of the processing being done... how would
I go about producing, or otherwise obtaining, a file which matches the
specs of the one you're testing with?

-- 
       The Wanderer

Warning: Simply because I argue an issue does not mean I agree with any
side of it.

Secrecy is the beginning of tyranny.