[mythtv-users] Re: Channel change - multiple ringbuffers

[Edward Wildgoose]

> FYI, I have an AMD 1200, and it takes about 3 seconds to change
> channels using EXT3. I'll change it to ReiserFS tonight and see if it
> helps.

Guys!! Isaac changed the code a week or so back in CVS in such as way that
deleting the ring buffer is highly unlikely to be a slow event anymore, even
on Ext3!  Dozens of people have reported that even changing to XFS makes no
significant difference!

Look, I'm really interested in speeding up channel changing as much as the
next person, but you need to look elsewhere than just banging on about the
ringbuffer deletion time.  It is quite likely that the code is "simply"
choosing to buffer quite a lot of data before starting playback.

Why not just drop into the code and put some print lines in around channel
changing code and do some benchmarking until you can figure out where the
time is actually spent.  Report *this* back to the list (or preferably do
something about it).  It's not too hard to do this kind of crude profiling
to see where the code is actually spending time and it means you can report
times back to the nearest millisec or whatever your PC timer resolution is.

OK, appeal over.... Off you go then...

Ed W