New Windows vs. Linux Benchmarks Are On The Way

[phoronix]

Phoronix: New Windows vs. Linux Benchmarks Are On The Way

A few months back we began publishing some open-source Windows benchmarks once the Phoronix Test Suite started picking up Microsoft Windows 7 x64 support to join Linux, FreeBSD, OpenSolaris, and Mac OS X as another supported testing platform. We had published Windows 7 vs. Ubuntu 10.04 benchmarks and then a benchmarking threesome of Windows, Ubuntu, and Mac OS X. We had also benchmarked EXT4 vs. the NTFS file-system while the real loser throughout all of this testing was the Intel Linux graphics driver as being the really sore spot for Linux performance when compared to Windows throughout all of our tests. A new set of Windows vs. Linux benchmarks are to be carried out over the next week or two and we will be running even more tests this time around...

http://www.phoronix.com/vr.php?view=ODQ0OQ

[peepingtom]

"To be clear, the only 2010 ThinkPad system with switchable graphics capabilities is the ThinkPad T410s. The ThinkPad T410 and T510 and W510 are either integrated graphics only or discrete graphics only. It might be possible to do a hack to allow “cold” switching between the integrated graphics core on the CPU or the discrete graphics chip, but this is not something we’ll support or even say is possible on any given system. There is extra circuitry required to enable switching capability. "

- Lenovo's Matt Kogut @ http://lenovoblogs.com/insidethebox/?p=406

It sure would be nice if I could donate money to the person who figures this out


Also, these laptops run UEFI in the lab: http://lenovoblogs.com/insidethebox/?p=560

[Linuxhippy]

So if the results are on the way - it would be great if you could wait with the news when the results are here

[tulcod]

So if the results are on the way - it would be great if you could wait with the news when the results are here

amen. it would be awesome if phoronix would start respecting their readers' time for once. publish when you've got something interesting, not when you haven't published for 4 hours. I don't mind waiting, updates roll in automatically thanks to my RSS readers.

[Michael]

amen. it would be awesome if phoronix would start respecting their readers' time for once. publish when you've got something interesting, not when you haven't published for 4 hours. I don't mind waiting, updates roll in automatically thanks to my RSS readers.

If you haven't paid attention to similar such threads, there's a reason notices like these are done in advance... Usually some readers have particular test requests or other feedback. However, when the testing is already complete and published, it's usually too late or inconvenient for me to go back and run such tests.

[miki100]

Please, include openSUSE 11.3 in this test. I would like to see how my favourite distro would look here

[Michael]

Please, include openSUSE 11.3 in this test. I would like to see how my favourite distro would look here

OpenSUSE 11.3 doesn't run much differently from Ubuntu, etc... i.e. http://www.phoronix.com/vr.php?view=15119

[mikeplus64]

OpenSUSE 11.3 doesn't run much differently from Ubuntu, etc... i.e. http://www.phoronix.com/vr.php?view=15119

They differ a lot in package management, and this is the one thing that is /never/ "benchmarked" even though it is one of the most important things to compare distros with, most of the other tests usually end up being because of different compiler/graphcs driver/other-versions-of-things.

This is the main thing that I don't like about these distro comparisons, most of the time the results are because of different versions of some things or other very obscure factors ( eg. in the 'five-way' distribution comparison you couldn't get 3D acceleration working in Arch Linux, didn't specify any X.org version and other details that you did specify for the other distros ) that aren't really relevent to the distro itself - things like package management, the way that it boots ( with distros unlike Arch or Gentoo for instance that have a 'stock configuration' to test in the first place ) and the time taken on the other hand are actually useful to test.