On 10/17/2015 03:16 PM, Fred Jan Kraan wrote:
On 2015-10-17 05:57 PM, Jonathan Wilkes wrote:
Fred Jan, Now that's interesting. Thanks for testing it.
What happens with larger data sets? One thousand, ten thousand, etc.
Same result for 10.000 items; first the bang, then the result from the dump. No interruption of the sound.
Now try that same patch on RPi and load five or ten files per second and chances are you will have xruns (order should remain the same, no matter what). Also, testing these in hypothetical/light cases where pd does nothing else of significance is unlikely to give you a real-world picture. FWIW, in L2Ork one of the biggest xrun culprits has been [coll] and data structures with large screen redraws that stole too much cpu away from audio processing thread, despite all the optimizations and prioritizations/lowlatency kernel/etc. Now, some of it is undoubtedly svg-enabled canvas and tkpath, some may be Ubuntu and its Unity desktop. Still, switching to threaded coll solved most of the problems. Another thing to consider, Max tests I shared earlier were performed on a fully loaded last gen MBP...
Best,
Ico