Last July we started writing about this,but never finished. It’s still not finished, but it is publicly available, and posted here: https://github.com/squinkylabs/SquinkyVCV/blob/master/docs/efficient-plugins.md
Thank you for putting this together - it is very clearly written and helpful!
Thank you - very useful!
One suggestion: you say “it is very easy to improve this using SSE or AVX instructions” as though they’re equivalently supported, but they aren’t.
If you’re going to mention AVX, you should probably also explain that Rack supports processors with SSE but not AVX. If you use AVX, you need to check at runtime that it’s supported, and fall back to SSE or plain old maths if not.
that’s a good idea. thanks!