forkrun VS jq

EDIT: moved code + example to GITHUB

One example that I dont know how to do streaming for involves breaking up the stdin stream and sending it to different places, but where you determine where to send it in realtime. This isnt the code that spawned this particular thread, but a good example of this is in this function I wrote called forkrun.

If you want an example of a really optimized bash script check out my forkrun utility. I spent a really long time optimizing it. It parallelizes loops (like xargs -P and parallel), and for problems with "many very quick iterations" (e.g., sha256sum of a million tiny files) it is twice as fast as the fastest method available via xargs or parallel (which means it is outpacing well optimized compiled C/C++ binaries).

as it so happens, I recently wrote a script called forkrun that parallelizes loops really fast. In particular for things like your simulation that require running 1 at a time in parallel (N parallel batches of 1 simulation, as opposed to N parallel batches of M simulations) the parallelization framework itself is 7x faster than xargs -l 1 -P N and ~25x faster than parallel -j N.

However, if you have bash 4 or later, using bash coprocs can offer much better performance. The idea is that you spawn a bunch of persistent coprocs and pipe data to them, allowing you to parallelize the loop without forking each individual iteration. I wrote / am writing a utility called forkrun that parallelizes loops using bash coprocs. You're welcome to use it, though heads up it is still in beta (almost everything works, but I am still debugging the -k flag sometimes not working properly...-k ensures the output order is the same as input order). If nothing else it gives a real world example of parallelizing using coprocs.

Side note: in case you were wondering the actual code that inspired this post is my forkrun utility. If you felt like checking it out Id love to hear any feedback you might have on it.

For me? My personal best script is (by a wide margin) my forkrun utility. It uses bash coprocs to efficiently parallelize loops. Usage is nearly identical to xargs -P <#>. Options are similar to xargs....a few are missing (namely those related to running things interactively), and a few options are things xargs doesnt do but forkrun can.

I wrote a nifty function for parallelizing loops in bash really fast called forkrun. It will, with the help of tail -F, let you do this (among many other things) by writing the output from each parallel worker to a separate file, which you can then monitor in real time with tail -F in another terminal window.

That should recursively list directories, counting only the files within each, and output² jsonl that can be further mangled within the shell². You could just as easily populate an associative array for further work, or $whatever. Unlike bash, zsh has reasonable behaviour around quoting and whitespace too.

¹ https://zsh.sourceforge.io/Doc/Release/User-Contributions.ht...

² https://github.com/jpmens/jo

³ https://github.com/stedolan/jq

What I miss from jq and what is implemented but unreleased is platform independent line delimiters.

jq on Windows produces \r\n terminated lines which can be annoying when used with Cygwin / MSYS2 / WSL. The '--binary' option to not convert line delimiters is one of those pending improvements.

https://github.com/stedolan/jq/commit/0dab2b18d73e561f511801...

If you have jq installed you can use it to make the output look nicer.

It requires jq for JSON processing and GNU parallel for concurrent searches in the notebooks.

jq is your friend.

The jq command (https://stedolan.github.io/jq/) is useful pulling that information out.