keithwood wrote:mystran wrote:No, I'm saying that assuming sufficient number of CPU cores to parallel process everything that can be parallel processed without additional latency, the longest (in sense of processing time) serial path that exists in the audio processing graph is still limited by single-threaded performance.
How many cores is sufficient? Let's say I'm doing MIDI mock ups of orchestral or film scores.
That's a question that's impossible to answer, because it depends a lot on exactly what your load distribution looks like. If you have a single expensive serial dependency chain and a bunch of cheap parallel tracks to go at the same time, you might even be fine with two and you're performance is almost completely limited by the single-threaded performance. If you have a lot of equally expensive tracks in parallel, then you might need up to one core per track to run it fully parallel and you can probably approach the total through-put limit of the system quite well whatever the number of cores.
In practice, most real-world audio work loads are somewhere between these extremes. In practice, your parallel opportunities are usually decent on track level, worse when you start working on the sends and basically none if you have to do master processing.