Multicore Audio Graphs



So far I have been working with SuperCollider, and they have the SuperNova Audio Server that takes advantage of multicoreprocessing, which is great, apart from the fact that it produces tons of audible drop outs. So with no experience in JUCE so far, my question is: Is it possible to have one AudioGraph processed by several cores? It kind of relates to this thread:

I was hoping to find some more up-to-date information on that topic.





I once did a proof of concept Multiprocessor version of the AudioProcessorGraph (included in JUCE)

My conclusions

- pre-sorting is unnecessary (for the order of processing, only for feedback detection), order should be calculated dynamically live

- use a hard spin-lock in the audio-callback to synchronize the threads (Juces Spin-Lock uses OS-Callbacks)