I'm posting this to discuss whether questions about performance optimizations are off-topic in data science, and of course to see better definitions for such questions. My main relation with data science -- which I'm not sure if it does actually exist right now -- is via performance improvement/distributed computing. Some of the questions I've posted, which I consider to be related to performance optimizations, are:
- How to speedup message passing between computing nodes;
- What is the difference between global and universal compression methods?;
- Why is it hard to grant efficiency while using libraries?;
- How to measure execution time on distributed system;
- How to compare experiments run over different infrastructures.
My point is: since these are not programming questions, and rather performance issues/techniques to help processing very large databases, they would (to me) be better fit to this site than StackOverflow, for example. Again, I'm not sure about what exactly would limit such questions, and neither if they are actually on topic here in DataScience.
Hope the answers shed light on a better definition for "performance optimization" questions, as well as decide whether they are on topic.