2010-09-27, 05:22
Now that the Dharma branch is underway are there any plans to deal with the high CPU use since the fork? I understand this was originally due to the xbox event processing loop but now that xbox support is going away are there plans to abstract this to platform-specific native runloops?
It would certainly lower the dharma cpu footprint, which for me on both my Mac and Linux boxes is about 20% at rest (vertical sync turned off, rss feed turned off).
I know it's peppered throughout the code and would be a pretty large undertaking but that nonblocking polling loop is the only reason I don't keep XBMC running all the time.
If this is in the roadmap I'll be glad to try and help out wherever I can.
Keep rockin' y'all
It would certainly lower the dharma cpu footprint, which for me on both my Mac and Linux boxes is about 20% at rest (vertical sync turned off, rss feed turned off).
I know it's peppered throughout the code and would be a pretty large undertaking but that nonblocking polling loop is the only reason I don't keep XBMC running all the time.
If this is in the roadmap I'll be glad to try and help out wherever I can.
Keep rockin' y'all
