Skip to content

Latest commit

 

History

History
161 lines (114 loc) · 5.38 KB

perf.md

File metadata and controls

161 lines (114 loc) · 5.38 KB
Raw

Benchmarks Comparing choose to Other Tools

"Results" section is generated from this script. It reports the task-clock for each command (which can differ from elapsed time). The script provides options regarding sorting and uniqueness. The defaults were used but different options lead to different results. The matrix of possibilities would be very large, so only the defaults are shown below.

Also note the compile time option called DISABLE_FIELD. It disables the --field arg, and removes some information associated with each token. This provides a small boost to sorting and uniqueness since there's a smaller memory footprint. --field is not disabled by default, and inline with above, the benchmarks are run on the default options.

Summary

Grepping

pcre2grep and choose have about the same speed.

Stream Editing

sed reads input until it reaches a newline character, and puts the content thus far in a buffer where it is then manipulated. Because of this, sed performs extremely poorly on input files that contain many small lines (for the no_duplicates case below, sed with a newline delimiter (default) is x50 to x100 slower than choose). To normalize the performance, the -z option was used with sed (to change the delimiter to a null char, which never occurs in the input). choose doesn't use delimiters in this way, and can't come across this type of pathological case. After this normalization, sed is faster than choose except in cases where there are few substitutions to apply.

Uniqueness

choose is faster than awk except in cases where there are few duplicates.

Sorting, and Sorting + Uniqueness

sort is using naive byte order (via LC_ALL=C), as this is the fastest. sort is faster than choose at sorting. If truncation is leveraged, or if there are many duplicates (when applying uniqueness as well), then choose is faster than sort.

Input Data

Each input file is the same size (50 million bytes), but the type of data is different.

plain_text

This file represents an average random workload, which includes text from a novel repeated.

The Project Gutenberg eBook of Pride and prejudice, by Jane Austen

This eBook is for the use of anyone anywhere in the United States and
most other parts of the world at no cost and with almost no restrictions
whatsoever. You may copy it, give it away or re-use it under the terms
...

test_repeated

This file has the line "test" repeated. "test" is the match target used throughout, below. This includes grepping for the word "test", or substituting "test" to "banana".

test
test
test
test
test
...

no_duplicates

For filtering by uniqueness, there are two extremes. One is where the entire file consists of the same element repeatedly, which is in test_repeated.txt. The other is when every element is different. This file counts upwards from 1 for each line:

1
2
3
4
5
...

csv_field

This file has a field where sorting and uniqueness should be applied. Note that even though the field is numeric, both sort and choose are using a lexicographical comparison in the benchmark.

garbage,1,garbage
garbage,2,garbage
garbage,3,garbage
garbage,4,garbage
garbage,5,garbage
...

Results

Versions

choose 0.3.0, ncurses 6.2.20200212, pcre2 10.43
pcre2grep version 10.43-DEV 2023-04-14
sed (GNU sed) 4.7
GNU Awk 5.0.1, API: 2.0 (GNU MPFR 4.0.2, GNU MP 6.2.0)
sort (GNU coreutils) 8.30

Specs

5.15.90.1-microsoft-standard-WSL2
AMD Ryzen 7 3800X 8-Core Processor
ram: 16331032 kB

Grepping

(ms) choose pcre2grep
plain_text 247.17 269.69
test_repeated 1620.34 1583.77
no_duplicates 323.59 370.16

Stream Editing

(ms) choose sed
plain_text 179.57 135.57
test_repeated 2725.50 1157.39
no_duplicates 5.10 44.00

(here is a cherry picked great case for choose compared to sed)

(ms) choose sed (with newline delimiter)
no_duplicates 5.13 543.93

(a special case, where choose cheats by using a literal replacement string)

(ms) choose (delimiter sub) sed
test_repeated 1521.93 1156.64

Sorting

(ms) choose sort
plain_text 1628.88 448.06
test_repeated 1850.89 1616.13
no_duplicates 3714.94 1036.93

(a special case that leverages truncation)

(ms) choose -s --out 5 sort | head -n 5
no_duplicates 354.20 1059.81

Uniqueness

(ms) choose awk
plain_text 111.95 214.41
test_repeated 565.31 1147.75
no_duplicates 2340.37 1496.42

Sorting and Uniqueness -u

(ms) choose sort
plain_text 122.80 440.57
test_repeated 558.86 1640.79
no_duplicates 5742.11 1168.84

Sorting and Uniqueness based on field -u

(ms) choose sort
csv_field 2770.27 474.02