Home > apple, computer, hackintosh, linux, macintosh, microsoft, windows > Leopard vs XP vs Vista vs Linux part 2

Leopard vs XP vs Vista vs Linux part 2

operating systems challengeIn part 1 I looked at results from Geekbench, Photoshop and VLC to see how much of an impact the operating system has on application speed. For part 2 I used two more cross platform test, the first person shooter Doom 3 and 3D modeller and renderer Blender.

Doom 3 was released in August 2004 for Windows and ported to Linux. A Macintosh version followed a few months later, and all three versions have been updated through various patches. I tried the game when it was first released but my Athlon XP 2600 and Nvidia 6600GT struggled at anything other that low resolution and the lowest quality settings. For this benchmark I installed the full version on Mac/XP/Vista/Linux and added all the updates available. That was definitely necessary, as the original Mac version was PowerPC only and ran very slowly under Rosetta. In the game’s preferences, quality was set to ultra and the resolution to 1280×1024. From the console I used ‘timedemo demo1’ to run a pre-recorded sequence, and the results are the average frame rate per second given at the end of this.

Platform
Mac
XP
Vista
Linux
timedemo demo1
ultra quality at 1280×1024
42.2
49
44.2
32.7

No real surprise that XP was the clear winner here, since it’s been available for several years and Nvidia have had plenty of time to optimize the drivers. Vista is widely reported as being slower running games than XP, but the result wasn’t as bad as I expected from the negative publicity. The Mac result was surprisingly competitive considering you can’t download the latest drivers from Nvidia and are stuck with what Apple provide. Linux Mint didn’t fare too well, though it was using the drivers installed by the control panel restricted drivers option. I used the Envy software package to install the latest Nvidia drivers for the 7300GT card, which made no difference to the results. To see if other distributions are any better I installed Ubuntu 7.10, OpenSUSE 10.3 and Fedora 8. Ubuntu was the only one I could get working with Doom 3, and the results were exactly the same as Mint. The OpenSUSE installation went horribly wrong so didn’t get tested (more of that in the next article) and I couldn’t get Fedora to run the game due to an audio problem.

Blender was next, and I used installers from the blender.org site for Mac and Windows. Linux Blender was installed through the software portal. For the test I downloaded the test.blend file from www.eofw.org/bench/ which made it very easy to run a standard benchmark. I didn’t tweak or optimise any settings to ensure the same results anyone would get after installing the software.

Blender test.blend
Mac
XP
Vista
Linux Mint
Ubuntu
OpenSUSE
1 thread
1m38s
1m57s
1m57s
2m13s
2m12s
2m16s
4 threads
28s
33s
32s
36s
36s
38s
8 threads
26s
31s
32s
35s
35s
36s

Again some interesting results. The one thread test used one CPU core, the four thread test used all four cores and the eight thread test used every bit of power available with two threads per core. After reading the Blender Linux FAQ I expected to see linux give faster results than the other operating systems, but it came last on all three of the distributions I tried. Mac OS knocked five seconds off XP’s time in the four and eight thread tests, and although the single thread test was nineteen seconds faster there wouldn’t be any point in using it in a real world situation.

The obvious question at the end of these tests is which operating system you should use. The advice after these far from conclusive tests is the same that’s been given in countless reviews. Windows XP gives the best performance for games, has massive market support and is reliable. Vista is getting better, file operations are much improved under service pack one, and it’s quite pleasant to use once you get used to the changes. Linux is improving in huge steps and has come a long way since I first bought SUSE 6.0 many years ago. Linux Mint is the most impressive distribution I have tried so far, and the video/audio codec support as standard is a welcome addition. But it’s still not as easy to use as Windows or the Mac OS, and some of the performance claims I have seen are no longer true. Mac OS is also improving fast, and Leopard is still my choice of this group. Graphical performance is not too far behind windows, and in all the other tests Leopard performed best or close to best. It also feels the most natural to me, and that’s the most important aspect than no benchmark can capture.

Advertisements
  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: