Open Benchmarking: More Windows Than Linux

[phoronix]

Phoronix: Open Benchmarking: More Windows Than Linux

Earlier this week I posted some statistics about the increasing rate of Linux news and some OS/browser stats for Phoronix. As many readers found it interesting, here's some stats for OpenBenchmarking.org...

http://www.phoronix.com/scan.php?page=news_item&px=OB-Windows-Linux-Stats

[timofonic]

- More extreme file system benchmarks, including different distributions and all kind different file system configurations.

- Include advanced power usage monitoring on all systems, fine grained if possible. You can use Sigrok plus custom hardware or use available ones, there's nice electronics geeks out there in the OSHW world.

- Adding more embedded systems running different Operating Systems if possible. ARM, MIPS, Coldfire, PowerPC, Blackfin, FPGA platforms....

- OpenCL and CUDA tests.

- Very detailed Windows vs Linux vs MacOS X vs BSD performance benchmark analysis.

- Extreme stress tests to BtrFS versus other file systems (XFS, EXT4, Reiser4, ZFS...) to check reliability and performance, but under *all* kind of circumstances.

- Wine vs native (Linux, Windows, Mac) games performance.

- Advanced OpenGL tests of newest MESA features vs propietary drivers.
* Some way to automatically check performance issues of software using OpenGL? Games, Blender, KiCad...

[Michael]

- More extreme file system benchmarks, including different distributions and all kind different file system configurations.

- Include advanced power usage monitoring on all systems, fine grained if possible. You can use Sigrok plus custom hardware or use available ones, there's nice electronics geeks out there in the OSHW world.

- Adding more embedded systems running different Operating Systems if possible. ARM, MIPS, Coldfire, PowerPC, Blackfin, FPGA platforms....

- OpenCL and CUDA tests.

- Very detailed Windows vs Linux vs MacOS X vs BSD performance benchmark analysis.

- Extreme stress tests to BtrFS versus other file systems (XFS, EXT4, Reiser4, ZFS...) to check reliability and performance, but under *all* kind of circumstances.

- Wine vs native (Linux, Windows, Mac) games performance.

- Advanced OpenGL tests of newest MESA features vs propietary drivers.
* Some way to automatically check performance issues of software using OpenGL? Games, Blender, KiCad...

Those seem like all test requests you have for articles on Phoronix, not OpenBenchmarking.org features.

However, most of them are simply not feasible without a lot more subscribers/donations. E.g. only affording one power meter as it is right now, would need many more to cover all tests I run. Also no access to lots of the embedded systems, etc.

[Yorgos]

It's like the articles,
more biased than objective.

[GreatEmerald]

For OpenBenchmarking improvements, did you check out the state of the "OPC Classification"? For me it looks broken, text is on top of other text. It's also very unclear what it all means. What are the background colours, and what does the line position mean? "Percentile rank" is super vague.

[mitcoes]

I would add WINE (and MS WOS) opengl, directx9, 10 and 11 plus future Vulkan game benchmarks as there are so many MS WOS users if some day WINE performance is good enough for them perhaps they will switch to GNU/Linux. And a WINE / MS WOS score and overall score would be a great benchmark to read.

[Michael]

For OpenBenchmarking improvements, did you check out the state of the "OPC Classification"? For me it looks broken, text is on top of other text. It's also very unclear what it all means. What are the background colours, and what does the line position mean? "Percentile rank" is super vague.

Whoops, looks like a regression in that part of the graph code. Will get it cleared up.