Ubuntu 7.10 vs. 8.04 Benchmarks

Phoronix: Ubuntu 7.10 vs. 8.04 Benchmarks

Back in December we looked at the initial Ubuntu 8.04 Alpha 2 performance by comparing it to Ubuntu 7.10. In that article we had found the performance between the two releases to be roughly the same. Now that we're nearing an end in this development cycle as Ubuntu 8.04 LTS (Hardy Heron) will be released later this month, we've ran a new set of benchmarks comparing the latest Ubuntu 8.04 packages to the previous Gutsy Gibbon release.

http://www.phoronix.com/vr.php?view=12181

It is basically impossible to gain more benchmark speed with an os upgrade. You can only "feel" it faster when some newer kernel options are used. But those are not visable in your benchmark. What I dislike a bit about the latest Ubuntu releases is that there is no live option to disable compiz. You have to use vesa mode if you don't want it. Compiz is not always fun, it adds some effects (which are very restricted in default settings) but video playback does not really work optimal. I prefer such kind of eye candy optional - 7.10 was even fully unusable due to a bug in mesa for my X700SE, now there is a patch in mesa which fixes this problem.

If compiz "cannot" be disabled, then it is a major flaw for me. On my system (F8), it makes the nvidia proprietary driver somewhat unstable in demanding situations, and it decreases gaming performance. So it is useful to be able to disable it.

(of course, with Fedora 9, out of the box Compiz support is very bad: the ugly Fedora "desktop effects" box is still there)

It is possible later but when you think of the desktop image as live cd then you have a problem in case of driver issues.

It amazes me how useless the phoronix benchmarks usually are. I can tell you from my years of linux on the desktop usage that hardy will be a big upgrade for responsiveness and interactivity because of the new CFS scheduler. There is benchmarks for that kind of thing too, but they are just as dull. Basically multitasking is much improved. This is something most people that are power users or people that use computers daily for desktop stuff can notice quickly.

I tested it myself about a month ago and it was DOUBLE the Gutsy memory usage. Please do some official tests.
Thanks

Boot time and memory usage are valid benchmarks, I think, but a few gaming benches thrown in to spot major "accidents" (like we can see with Vista ...) still seem like a good idea to me.

I think many users do not count memory usage correctly.

free -m|awk '/Mem/{print $3-$6-$7}'

Should be the correct value.

Actually the gnome system monitor reports the same numbers as your terminal command, for "user memory".

Wasn't really expecting speedsups on 8.04, but with all of the new stuff in, I'm glad they kept the good pace :)

I guess when you install only the kernel with gutsy you get the same feeling of speed increase. For Kanotix I also use (slightly modified) hardy kernels, some patches even came from me.