2014-02-25, 19:37
I am trying out a new somewhat lightweight skin (this one).
It is running fast and nice…. in the beginning, I noticed that performance gets worse and worse for every minute (small explanation here).
This even if I leave the Pi alone.
Of course I suspected the slide show etc. in the background, but wasn´t sure that this is a bug (caching etc.) or if this is not something that Pi can handle.
I suspected this PR and I am now running a build prior to that to see if there is an improvement.
It seems to work much better.
@popcornmix
Do you think that this is a regression in PR#4234 or is Pi or another commit since the 13th?
Running build from 13th Feb (priot to PR#4234) and everything is much better. A bit early yet though, but it is a lot faster so far, running for almost 30 minutes which before showed a big regression (almost unusable).
Should I make a new build with that PR reverted to see if it works better or do you have any other idea´s?
Complete log from latest build (regression): http://www.xbmclogs.com/show.php?id=138634
Complete log from 13th FEB build (working): http://www.xbmclogs.com/show.php?id=138637
Edit: I think I start to see small regression on 13th build as well, but not as fast and not as big. I´ll run it for a short while and see if it really is better, and if so how much.
It is running fast and nice…. in the beginning, I noticed that performance gets worse and worse for every minute (small explanation here).
This even if I leave the Pi alone.
Of course I suspected the slide show etc. in the background, but wasn´t sure that this is a bug (caching etc.) or if this is not something that Pi can handle.
I suspected this PR and I am now running a build prior to that to see if there is an improvement.
It seems to work much better.
@popcornmix
Do you think that this is a regression in PR#4234 or is Pi or another commit since the 13th?
Running build from 13th Feb (priot to PR#4234) and everything is much better. A bit early yet though, but it is a lot faster so far, running for almost 30 minutes which before showed a big regression (almost unusable).
Should I make a new build with that PR reverted to see if it works better or do you have any other idea´s?
Complete log from latest build (regression): http://www.xbmclogs.com/show.php?id=138634
Complete log from 13th FEB build (working): http://www.xbmclogs.com/show.php?id=138637
Edit: I think I start to see small regression on 13th build as well, but not as fast and not as big. I´ll run it for a short while and see if it really is better, and if so how much.