Ubuntu 8.10 significantly slower than previous versions
Pau Garcia i Quiles
pgquiles at elpauer.org
Thu Nov 6 19:41:18 UTC 2008
Quoting Bryce Harrington <bryce at canonical.com>:
> On Thu, Nov 06, 2008 at 03:58:51PM +0100, mr wrote:
>> According to the recent benchmarking article by Phoronix, the previous two
>> releases of Ubuntu are significantly slower than Feisty Fawn. In some cases
>> this can be seen as up to 50% performance drop with certain desktop tasks.
>> I can confirm that this is true in that my girlfriends desktop used to be
>> quite capable of playing a 1080p x264 video but since upgrading to gutsy and
>> then hardy it has become unwatchable, even mplayer reports that "YOUR
>> COMPUTER IS TOO SLOW"
>> I think that the reasons behind this reduction in performance across the
>> board needs some serious investigation and work done to reverse this trend.
> Indeed, half the reason I suggested Phoronix do these tests is to
> stimulate more investigation into performance issues. We've had
> anecdotal evidence of performance reductions since Gutsy at least, and
> Phoronix presented a good opportunity to get some solid numbers.
> I've spoken with upstream about -intel performance previously. They've
> indicated their focus is on the current git version of the driver, and
> so would ask that anyone wishing to provide feedback on performance to
> first run the git-head version of the driver. This would enable the
> user to update and give swift feedback to the developers on any
> performance changes they are experimenting with.
> Beyond that, I'd encourage anyone wishing to help improve -intel
> performance on Ubuntu to join the ubuntu-x mailing list to discuss it in
> additional detail.
The disk IO performance decrease from Gutsy to Hardy is anything but
) bug report in the Zumastor project has some (quite scarce) info,
although if you read the IRC logs, you'll realize how much pain it
caused (it made Zumastor unusable due to slowness). It seems that the
scheduler included by default since kernel 2.6.23 cuts performance in
half for certain IO operations.
Same thing for the CPU scheduler. For instance, here VMware Server is
barely usable when high disk IO is requied because for some reason,
after a few seconds of good performance (which implies high CPU usage)
the kernel starts throttling CPU and VMware starves for a few seconds.
What surprises me about this CPU throttling is it achieves nothing:
VMware is left with 10% of CPU and the other 90% is idling!
I'm forward-porting kernels 2.6.20 (Feisty's) and 2.6.22 (Gutsy's) to
Hardy to verify it is the scheduler what is causing trouble. It will
be available in my PPA ( http://launchpad.net/~pgquiles/+archive )
tomorrow, in case anyone is interested.
Pau Garcia i Quiles
(Due to my workload, I may need 10 days to answer)
More information about the Ubuntu-devel-discuss