On Tuesday 19 August 2008 18:35:18 Håkon Alstadheim wrote:

> When I say "tune" I mean things like picking a resolution and a
> deinterlace method for video that is as good as possible, while still
> leaving enough headroom to avoid uneven playback. 

Ah, that puts a different spin on it. Those things are outside my area of 
expertise

> Given that a schedule
> update or a backup run might kick in at a bad time despite all efforts
> to write a good crontab, knowing when to stop is not always easy. I run
> all system tasks under "nice ionice -c3", and they will still cause
> hiccups if the system is maxed out. 

nice, ionice and all other tools that look like they might affect scheduling 
are actually just hints to the Linux kernel, which is free to completely 
ignore them (and very often does). In fact, for a great many years in the 
early days, nice actually did nothing of any consequence at all... :-)

nice exists because it is a long-standing Unix tool and on early Unixes it did 
do what the man page implies. Things have changed and kernel writers realise 
that.

re crons, there's not a lot you can do when vixie-cron decides to kick in. 
However, there are other cron daemons about. I've never used it, but I recall 
a discussion here a year ago about fcron which looks like it might suit your 
needs. IIRC you can configure it to delay a cron job while a specified process 
is running and all manner of other cool stuff suitable for desktops 

> mplayer kept saying my system was too slow, while the cpu was idling at
> 20%. Turns out top was correct in that instance, mplayer was
> misinterpreting input data, trying to play back at half the intended
> frame-rate. Now I'm past that hurdle and adding deinterlacing and other
> filters. I'll just have to hold back on the temptation to go all out on
> the filter options :-)

I find mplayer often gets it wrong, and the console messages about audio 
drivers being responsible for slowdowns are often true. I also usually find 
that a slow machine that has a light load is often due to I/O blocking. Have 
you looked into this yet?

-- 
alan dot mckinnon at gmail dot com

Reply via email to