# Why I stuck with Windows for 6 years while developing Discourse

DevFeed: [Why I stuck with Windows for 6 years while developing Discourse](<https://devfeed.tech/articles/why-i-stuck-with-windows-for-6-years-while-developing-discourse-22432.md>)

Original publisher: [Read original article](<https://samsaffron.com/archive/2019/03/31/why-i-stuck-with-windows-for-6-years-while-developing-discourse>)

Author: Sam Saffron

Published: 2019-03-31T23:44:35Z

Content type: article

Language: en

Sources: [Sam Saffron](<https://devfeed.tech/sources/sam-saffron.md>)

Topics: [Windows](<https://devfeed.tech/topics/windows.md>), [Development](<https://devfeed.tech/topics/development.md>), [Ruby](<https://devfeed.tech/topics/ruby.md>), [Windows Subsystem for Linux](<https://devfeed.tech/topics/wsl.md>), [virtualization](<https://devfeed.tech/topics/virtualization.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Docker](<https://devfeed.tech/topics/docker.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [blog](<https://devfeed.tech/tags/blog.md>), [c-sharp](<https://devfeed.tech/tags/c-sharp.md>), [development](<https://devfeed.tech/tags/development.md>), [docker](<https://devfeed.tech/tags/docker.md>), [go](<https://devfeed.tech/tags/go.md>), [linux](<https://devfeed.tech/tags/linux.md>), [net](<https://devfeed.tech/tags/net.md>), [performance](<https://devfeed.tech/tags/performance.md>), [rails](<https://devfeed.tech/tags/rails.md>), [ruby](<https://devfeed.tech/tags/ruby.md>), [rust](<https://devfeed.tech/tags/rust.md>), [saffron](<https://devfeed.tech/tags/saffron.md>), [sam](<https://devfeed.tech/tags/sam.md>), [sql](<https://devfeed.tech/tags/sql.md>), [virtualization](<https://devfeed.tech/tags/virtualization.md>), [weblog](<https://devfeed.tech/tags/weblog.md>), [windows](<https://devfeed.tech/tags/windows.md>), [wsl](<https://devfeed.tech/tags/wsl.md>)

## AI overview

The author examines the performance cost of developing Discourse on Windows while using WSL or a VMware-based Linux virtual machine. On the author's machine, running specs was 25% slower with WSL and 38% slower with VMware, while WSL also performed poorly with many small files.

## Source excerpt

I made this tweet that got reasonably popular: We benchmarked how long it takes to run the Ruby test suite for Discourse across our various dev machines. I can not believe what a crazy tax I have paid over the years insisting on sticking with Windows, highlighted results mine. image824x706 73.7 KB https://twitter.com/samsaffron/status/1111511735851081728 This evoked a lot of extremely strong emotions from various people out there. Ranging from "Sam is a fool what kind of insane benchmark is this", "the real story is MacOS has bad Ruby perf" to a general "Oh no". The core point I was trying to make was that I was paying a pretty high tax for deciding to "stick with with Windows". There are a bunch of other points hiding here that are also worth discussing. Why are you sticking with Windows to run Linux in a VM? https://twitter.com/shanselman/status/1111538525017530371 What I did not know is the extent of the VM tax I was paying regularly. I never dual booted my computer so I had no proper anchoring point of reference. I very strongly believe that many Ruby/Rust/Go/Elixir/Scala and even some Node developers who end up doing the WSL dance or run Linux in a VM for development, or use Linux Docker for dev on Windows are not aware of the full extent of the tax. On my machine the price of admission for using WSL was 25% slowdown in my day to day running of specs. And a 38% slowdown for using a VMware based VM. I am not alone here... other team members have experienced similar slowdowns. Other people out there also experience similar slowdowns. https://twitter.com/stevedesmond_ca/status/1111585969650712578 What I thought was inside my wonderful wish hat was that the performance hit was minor: serverfault.com Is virtual machine slower than the underlying physical machine? virtualization, performance, cloud-computing, benchmark answered by TomTom on 07:25AM - 24 Apr 10 UTC Yes. But that is not the question. The difference is normally negligible (1% to 5%). If you Google, well t