I've written a ChucK script to concurrently instantiate samples with their own effects chains with the aim of creating a kind of kaleidoscopic music box that randomly plays filtered longish samples (say up to two minutes). The box bit is intended to be a raspberry pi.
Running on a pi the program maxes out at two concurrent samples and a modest effects chain for each. Ideally I'd prefer to double or even triple that so clearly I need to scale back my resource-intensive intentions, along with attempting to improve the pi's performance (overclocking, for example). There are concrete steps to take on both these fronts, but I have two concerns I'd be keen to get advice on: