Skip to content

Instantly share code, notes, and snippets.

@jperkin
Created November 14, 2025 15:12
Show Gist options
  • Select an option

  • Save jperkin/a1065caaf09a4b31dea61f35005e389d to your computer and use it in GitHub Desktop.

Select an option

Save jperkin/a1065caaf09a4b31dea61f35005e389d to your computer and use it in GitHub Desktop.

While testing latest bindfs I wanted to see if there were any differences, so:

  • Create macos-trunk-arm64 sandbox
  • Build sysutils/coreutils
  • Time 'bmake' after depends has done.

Results:

original bindfs 1.17.7 --multithreaded

real    1m41.541s
user    1m15.336s
sys     1m16.597s

bindfs-1.18.3 --multithreaded

real    1m46.467s
user    1m16.200s
sys     1m23.400s

bindfs-1.18.3 not multithreaded

real    1m45.581s
user    1m15.562s
sys     1m21.857s

I was also curious to remember why I needed bindfs instead of using NFS mounts. To get /System mounted (which is the hot path due to lookups), I needed to also mount Preboot and Data, but after that was done the perf (despite nfsd being lower in top(1) while I was measuring compared to bindfs) was worse:

real    1m49.257s
user    1m12.550s
sys     1m39.673s
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment