In several examples (both in the VisualDSP install, and online training module), I've seen examples that reuse VDK messages, e.g.:
Thread1 creates messages at init time, and send them on a channel to thread2. Thread2 reads them, and send them back to thread1, which reuses them.
What is the tradeoffs between this method, and the simple (and logical method of):
- Thread1 creates a message when needed, and sends to thread2
- Thread2 pends on a message, and when receives is, gets its payload and then destroys it.
Obviously, method 2 is much easier to implement.
What method is to be used when writing complex applications?