What was tested
Gloor runs Windows 11 and CachyOS on the same desktop. He is upfront that this is "far from a scientific test", and the setup explains why.
Where Windows had the edge:
- Windows 11 sits on an NVMe drive. The tower has only one NVMe slot, so CachyOS lives on a SATA SSD. Gloor cites online benchmark databases putting the SATA drive at roughly half the write speed and about a fifth of the read speed.
- He disabled Copilot on Windows so the comparison would be closer to an AI-assistant-free Linux install. His reasoning is that this frees some resources, but he did not measure the effect.
- Windows had Fast Startup enabled for the boot test.
Where Linux had the edge:
- CachyOS was running its own optimized stable kernel. CachyOS's documentation describes that default kernel as built with Clang ThinLTO and profiled with AutoFDO, with a 1000 Hz timer tick. Those are real customizations. The documentation doesn't say how much they speed up any particular task, so they can't be credited for the results below.
- The desktop was Xfce on an X11 session, which is among the lighter environments you can pick.
Both installs were fairly clean. CachyOS was a few days old, and Windows had only a handful of basic apps. Neither was heavily bloated or tuned.
The numbers
| Test | Windows 11 | CachyOS (Xfce) |
|---|---|---|
| Bootloader to login screen | 19 s | 20 s |
| Idle RAM use | 6 GB of 12 GB | 1.5 GB (Btop++) |
| Launch Chrome and load Google's home page | 3.16 s | 1.88 s |
| Launch fully updated Steam | 32.94 s | 22.87 s |
Boot time: a wash
Windows reached the login screen in 19 seconds and CachyOS in 20. That is effectively a tie. Gloor argues it favors Linux because CachyOS was on the slower drive. That is a fair point, but it is an argument about hardware, not a measurement.
Idle RAM: interesting, not a speed test
Windows idled at about half of the 12 GB installed. CachyOS with Xfce showed 1.5 GB in Btop++. Gloor says the figure rises to about 3.5 GB if caching is counted, but he doesn't explain how he got that number. Switching to the experimental COSMIC desktop added only about half a gigabyte.
Gloor himself notes that idle memory says little about speed, because a modern OS puts spare RAM to work caching data. Low idle usage means more headroom on a small machine. It doesn't make a PC faster on its own.
Chrome and Steam: Linux ahead
Chrome opened in 3.16 seconds on Windows and 1.88 seconds on CachyOS. Steam took 32.94 seconds on Windows and 22.87 seconds on CachyOS. Those are real gaps, and they held even though Linux was on the slower disk.
Packaging matters here. Gloor used Chrome from the AUR and Steam from Arch's multilib repository. He warns that Flatpak or Snap builds may start more slowly. Anyone copying this test with a sandboxed package could see different results.
What the results do and don't show
The article reports single stopwatch timings. It gives no repeated runs, no variance, and no distinction between cold and warm starts. So the fair reading is narrow: on this machine, CachyOS/Xfce opened two apps faster than Windows 11, and boot times were close.
Several limits apply:
- It doesn't show that Linux is faster than Windows in general. This is one distribution, one lightweight desktop, and one tuned kernel.
- Storage was mismatched. That favored Windows, but nobody can say by how much for each task.
- The results aren't tied to any cause. Maybe Linux's lighter background load helped, or the app packaging, or the Windows-side services, or something else. The test can't separate these.
- Gloor's own verdict is more modest than the headline. He says he doesn't notice much difference in daily use on this desktop, especially after changing several Windows 11 settings.
How this compares with formal benchmarks
Gloor points to Phoronix's Ubuntu versus Windows 11 testing as the more rigorous evidence. That work measures something different. Phoronix put the geometric mean of more than 110 benchmarks at about 9% in Ubuntu's favor. The machine was an HP Z6 G5 A workstation, and Windows was HP's shipped Windows 11 Pro image.
The detail matters more than the headline. In those results:
- Phoronix says thread-happy CPU rendering is where Linux most often beats Windows 11 by a wide margin.
- GPU results were mixed. Windows 11 Pro had a small advantage in Vulkan ray tracing.
- Phoronix found Llama.cpp Vulkan AI performance broadly similar, with some Windows wins on CPU BLAS tests.
- For H.265 encoding with Kvazaar, Ubuntu was much faster.
- Running Ubuntu under WSL2 on Windows 11 carried a measurable cost for code compilation compared with bare-metal Linux. That may matter to developers.
Phoronix saw a similar margin on a different system. On an HP Z4 G6i with an Intel Xeon 678X, Ubuntu 26.04 LTS was around 9% faster overall. A WindowsForum report on that test noted that HP's Windows image used NVIDIA's R582 driver branch while Ubuntu defaulted to the newer R595 branch. That is a reminder that even rigorous OS comparisons carry configuration differences.
Neither Phoronix result is a stopwatch test of app launches on a modest desktop. They are workstation workloads on high-end hardware. They fit Gloor's results loosely at best. Both suggest Linux can do well, but neither shows a fixed percentage gain for everyday use.
Older hardware
Gloor says the biggest difference he has seen is on a 2011 Toshiba laptop. In his account it was unusable with Windows 10. After he installed Linux and tuned KDE Plasma, it was slow but usable. He gives no benchmark data, so treat this as one person's experience. It is plausible, though, since lighter desktops suit old machines with little RAM.
Practical takeaways
If you're curious whether a switch would help you, test the setup you would actually use.
- Match the storage. Put both systems on equivalent drives, or at least keep the mismatch in mind when reading results.
- Use the same kind of app package. Compare native installs with native installs, or Flatpak with Flatpak.
- Repeat each test. Run it several times and separate cold starts from warm ones.
- Compare the desktop you'd really run. Xfce numbers won't predict how GNOME or KDE Plasma behave.
- Test your own workload. Games, creative tools and anything that depends on drivers are where the platforms differ most. Gloor didn't test gaming and says Proton claims are anecdotal and need more testing.
- Tune Windows too. Gloor says his Windows impressions improved after he disabled several 11 settings. A fair comparison trims both sides.
Bottom line
On this one dual-boot desktop, CachyOS with Xfce opened Chrome and Steam noticeably faster than Windows 11, boot times were within a second of each other, and idle memory use was far lower on Linux. Gloor still didn't feel much difference in daily use. The headline question has a qualified answer: sometimes, and by margins that depend on hardware, desktop and workload. For most people on modern hardware, other factors will probably matter more than raw speed. Those include software compatibility, driver support and personal preference.