[mythtv] Effect of Multi-rec on channel change times....?
lists at wildgooses.com
Thu Jul 19 15:47:15 UTC 2007
> I don't know. All STB's I've used had pretty darn awful
> channel changing speed considering that they don't need
> to do the buffering that a PVR does.
The channel changing moan has come up so many times it would be useful
to add some instrumentation to the code to track it I think. I did hack
in some logging a while back on my own tree and found that there was
quite a lot of room for improvement, but really it needs to get some
proper instrumentation in the code so that:
a) less experienced myth video users such as myself can more easily see
the code flow during channel change (I found it hard to see how it jumps
around various threads)
b) it stops people moaning about areas of slowdown that quite clearly
aren't an issue and instead helps optimise the areas which will help
It's quite a long time back now, but I found about 3/4 of a second
disappearing in waiting on locks (which looked like it could be
removed), but I struggled to figure out which threads were actually
causing the holdup.
Could someone who knows the code better (Daniel?) please consider adding
in a dozen or so debug logging statements so that we can more easily see
where the bulk of the time goes in channel changing? I think there are
some easy fixups to be had here.
More information about the mythtv-dev