3 ms·
The fact that even in the test example it took ripgrep 0.115 s with 97% CPU usage vs Krep's 0.106 s with 328% CPU usage is a testament of why ripgrep is still p
by alkh 1y ago
The fact that even in the test example it took ripgrep 0.115 s with 97% CPU usage vs Krep's 0.106 s with 328% CPU usage is a testament of why ripgrep is still preferable. Such a huge tradeoff in resource consumption for such a small speed gain is hard to justify
- burntsushi 1y agoripgrep author here. Like many grep tools, it also has significant performance cliffs that are very easy to hit: $ time rg -o -c '[A-Z]' subtitles2016-sample.en 51835353 real 4.391 user 4.344 sys 0.042 maxmem 923 MB faults 0 $ time krep-1.0.2 -E -o -c '[A-Z]' subtitles2016-sample.en ^C real 8:02.95 user 3:04:10.58 sys 4.203 maxmem 919 MB faults 0 ripgrep has cliffs too, but they tend to be rarer and/or still give decent perf.