100+ Benchmarks Between Clear Linux vs. Arch-Based Manjaro Linux - Summer 2019 Tests

[phoronix]

Phoronix: 100+ Benchmarks Between Clear Linux vs. Arch-Based Manjaro Linux - Summer 2019 Tests

For those wondering how Intel's performance-optimized Clear Linux is comparing to the Arch-based Manjaro Linux distribution, here are a number of benchmarks on the same Intel Core i7 8700K in seeing how these rolling-release distros are competing for summer 2019...

http://www.phoronix.com/scan.php?page=news_item&px=Clear-Linux-Manjaro-2019-Summer

[tildearrow]

Why Manjaroooooooo?... There's a bunch of True Arch installing scripts out there...

[tildearrow]

dav1d results

4K

Clear: 119.15fps
Manjaro: 111.48fps

1080p

Clear: 402.57fps
Manjaro: 376.10fps

[rabcor]

Why Manjaroooooooo?... There's a bunch of True Arch installing scripts out there...

Why not manjaro?

[davidbepo]

wow, manjaro holds up pretty well, im proud to use it

[davidbepo]

Why Manjaroooooooo?... There's a bunch of True Arch installing scripts out there...

bootsplash, other kernel patches, driver manager util, pamac package frontend, and much more....

[tildearrow]

Why not manjaro?

.....

Manjaro is frankenarch.

OK, let's be more specific because apparently people want Manjaro.

Manjaro is not Arch. Manjaro very likely recompiles packages with different flags resulting in performance that differs from Arch.

[bridgman]

I like the new graph... don't have any other useful feedback yet though, seems to do its job well.

[boxie]

Michael I really like the new graph. As for suggestions -

• Render as SVG with a direct link to the open benchmarking result (i.e the test name in the first column links to the test result itself)
• hover text (if you can in SVG?) with the actual results per row
• Adding the geometric or harmonic averages at the bottom for a summary, (with a little bit of padding to distinctly separate it from the body of results
• Does it render error bars?

[randomizer]

The chart is quite long so I think it makes sense to order it such that the most interesting results are near the top, and the immaterial noise is at the bottom. You could sort it simply by magnitude, but that would spatially separate results for the same application where the magnitude varies significantly depending on the specific benchmark (eg. Flexible IO Tester). You could also sort it by total magnitude for each application (or some weighted metric to avoid skewing towards applications with many results of moderate magnitude), and then by each benchmark for a given application. This will keep results for applications like Flexible IO Tester together, but ensure that it sits below NAS Parallel Benchmarks.