With free-threaded getting closer to production ready, and a side-project that gets a bit of a speed-boost from real threads vs multiprocessing, I'm finally playing around with samply per py-free-threading.github.io/pr

It's pretty decent. Unfortunately it's telling me the bottlenecks are largely the ones I already know about โ€‹๐Ÿ˜‚. But it's nice to see the modern tooling with much more intuitive visualization back that up effectively.

Screenshot of "Samply" showing the flame-graph breaking down where samples showed the call stack of a python project were located proportionally
0

If you have a fediverse account, you can quote this note from your own instance. Search https://toot.cafe/users/atrus/statuses/115680934147114899 on your instance and quote it. (Note that quoting is not supported in Mastodon.)