New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
benchmark ripgrep on Windows #555
Comments
On Ubuntu 16.04 (ext4) rg is faster:
Probably it is due the relative slowness of NTFS on Windows.
It might not be the most honest comparison as this tests filesystem speed in Bash on Ubuntu on Windows instead of plain Windows.
https://github.com/msysgit/msysgit/wiki/Diagnosing-why-Git-is-so-slow Can you check if this git setting is enabled?
|
I still haven't had a chance to look into this, but I figured I'd respond with some things. First and foremost, the search times here appear to be in the lowish milliseconds, which means that noise could be a serious contributing factor. If one tool takes 250ms to do something and another takes 200ms to do the same thing, then they're pretty darn close and I wouldn't really be too concerned about that. Secondly, I've never done any benchmarks on Windows, and there are definitely things inside ripgrep that are specifically slower than on Unix systems, mostly having to do with the fact that paths in Windows are UTF-16 and all the glob matching logic requires UTF-8. Thirdly, @ghuls idea of looking to the file system may be a good idea because I think the appropriate action item here is that someone needs to do a comprehensive benchmark analysis of ripgrep on Windows, probably compared to other tools, and figure out which things are slow. I suspect the answers will be different than on Unix. |
I'd like to change the focus of this ticket from "ripgrep is slower than git grep" to "do comprehensive benchmarks of ripgrep on Windows." I don't know if I'll ever do that myself, but I think it is a worthwhile task. Also, the very first thing someone should do is find a meatier repository. |
I'm going to close this because I don't think there's a lot of value in tracking it. I'd still like for someone to do benchmarks with a similar rigor level that can be found in my blog post introducing ripgrep though! |
Per request from Twitter: https://twitter.com/burntsushi5/status/885481336261226496
My original statement was based on a few searches across my company's internal codebase, which I cannot share. However I was able to repro same thing using public code, so I wrote up details & results here.
Environment
Setup
Clone latest master from https://github.com/microsoft/visualfsharp (commit
b4c4d622135f8eb600cf032c6
)All snippets below are run in powershell.
Test 1
Basic search for
let
bindings in .fs files.Test 2
Basic search for
async
blocksThe text was updated successfully, but these errors were encountered: