jwalk VS dust

No, it's long known that parallel tree walking is able to process files/folders at better rates, especially for higher-throughput storages or even better, for hybrid or fully SSD storages. There are always some overheads on processing logic and there is no reason to wait for processing to finish while fetching another structure from disk, especially on high IOPS devices where single-thread programs are too slow. Since it's always better to refer to factual numbers, for instance comparison of Rust tree walkers, single-thread vs multiple-treads.

Try using dust https://github.com/bootandy/dust

Folks open to using gdu might like to try dust (6k stars), or even erdtree (1.4k stars) which is too recent to show up on lists like this and still a bit behind on stars. A lot of people seem to use starship (33k stars) though I'm personally oldschool on prompts. There are many other items on that list I'm not motivated to check.

The first thought I had was to use rayon for this. And looking at some prior art that does pretty much the exact same thing, it does indeed use rayon.

The dust CLI command (made with Rust) can do this too.

How does this compare to dust?

Happened to me this morning, something filled up my drive in minutes. I used dust[1] to look for large files while it was happening but knowing what was doing it would've been a big help.

[1] https://github.com/bootandy/dust

dust is also really nice.