Robot
/
fd
mirror of https://github.com/sharkdp/fd.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update benchmark results

This commit is contained in:
David Peter 2017-09-09 18:29:11 +02:00 committed by sharkdp
parent b34716051a
commit d67b1bba61
1 changed files with 54 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

75
README.md
View File

@ -9,10 +9,11 @@ While it does not seek to mirror all of *find*'s powerful functionality, it prov
## Features
* Convenient syntax: `fd PATTERN` instead of `find -iname '*PATTERN*'`.
* Colorized terminal output (similar to *ls*).
* It's *fast* (see benchmarks below).
* Smart case: the search is case-insensitive by default. It switches to
case-sensitive if the pattern contains an uppercase
character[\*](http://vimdoc.sourceforge.net/htmldoc/options.html#'smartcase').
* Colorized terminal output (similar to *ls*).
* Ignores hidden directories and files, by default.
* Ignores patterns from your `.gitignore`, by default.
* Regular expressions.
@ -35,31 +36,63 @@ complete (and more colorful) variants, see
[here](https://github.com/trapd00r/LS_COLORS).
## Benchmark
A search in my home folder with ~150.000 subdirectories and ~1M files. The given options for
`fd` are needed for a fair comparison (otherwise `fd` is even faster by a factor of 5 because it
does not have to search hidden and ignored paths):
Let's search my home folder for files that end in `[0-9].jpg`. It contains ~150.000
subdirectories and about a million files. For averaging and statistical analysis, I'm using
[bench](https://github.com/Gabriel439/bench). All benchmarks are performed for a "warm
cache". Results for a cold cache are similar.
Let's start with `find`:
```
benchmarking bench/fd --hidden --no-ignore --full-path '.*[0-9]\.jpg$' ~
time 2.800 s (2.722 s .. 2.895 s)
find ~ -iregex '.*[0-9]\.jpg$'
time 6.265 s (6.127 s .. NaN s)
1.000 R² (1.000 R² .. 1.000 R²)
mean 2.821 s (2.810 s .. 2.831 s)
std dev 16.52 ms (0.0 s .. 17.02 ms)
variance introduced by outliers: 19% (moderately inflated)
benchmarking bench/find ~ -iregex '.*[0-9]\.jpg$'
time 5.593 s (5.412 s .. 5.798 s)
1.000 R² (0.999 R² .. 1.000 R²)
mean 5.542 s (5.502 s .. 5.567 s)
std dev 37.32 ms (0.0 s .. 42.77 ms)
variance introduced by outliers: 19% (moderately inflated)
mean 6.162 s (6.140 s .. 6.181 s)
std dev 31.73 ms (0.0 s .. 33.48 ms)
```
(benchmarking tool: [bench](https://github.com/Gabriel439/bench))
Both tools found the exact same 14030 files. Note that we have used the `-iregex` option for `find`
in order for both tools to perform a regular expression search. Both tools are comparably fast if
`-iname '*[0-9].jpg'` is used for `find`.
`find` is much faster if it does not need to perform a regular-expression search:
```
find ~ -iname '*[0-9].jpg'
Concerning *fd*'s speed, the main credit goes to the `regex` and `ignore` crates that are also used in [ripgrep](https://github.com/BurntSushi/ripgrep) (check it out!).
time 2.866 s (2.754 s .. 2.964 s)
1.000 R² (0.999 R² .. 1.000 R²)
mean 2.860 s (2.834 s .. 2.875 s)
std dev 23.11 ms (0.0 s .. 25.09 ms)
```
Now let's try the same for `fd`. Note that `fd` *always* performs a regular expression
search. The options `--hidden` and `--no-ignore` are needed for a fair comparison,
otherwise `fd` does not have to traverse hidden folders and ignored paths (see below):
```
fd --hidden --no-ignore '.*[0-9]\.jpg$' ~
time 892.6 ms (839.0 ms .. 915.4 ms)
0.999 R² (0.997 R² .. 1.000 R²)
mean 871.2 ms (857.9 ms .. 881.3 ms)
std dev 15.50 ms (0.0 s .. 17.49 ms)
```
For this particular example, `fd` is approximately seven times faster than `find -iregex`
and about three times faster than `find -iname`. By the way, both tools found the exact
same 14030 files :smile:.
Finally, let's run `fd` without `--hidden` and `--no-ignore` (this can lead to different
search results, of course):
```
fd '[0-9]\.jpg$' ~
time 159.5 ms (155.8 ms .. 165.3 ms)
0.999 R² (0.996 R² .. 1.000 R²)
mean 158.7 ms (156.5 ms .. 161.6 ms)
std dev 3.263 ms (2.401 ms .. 4.298 ms)
```
**Note**: This is *one particular* benchmark on *one particular* machine. While I have
performed quite a lot of different tests (and found consistent results), things might
be different for you! I encourage everyone to try it out on their own.
Concerning *fd*'s speed, the main credit goes to the `regex` and `ignore` crates that are also used
in [ripgrep](https://github.com/BurntSushi/ripgrep) (check it out!).
## Install
With Rust's package manager [cargo](https://github.com/rust-lang/cargo), you can clone, build and install *fd* with a single command: