Speed on OpenSuSE versus other distro

I know this has been tackled many times before, but I’m looking for more information regarding the OpenSuSE system.

I just installed crunchbang linux and it’s very very zippy and fast. It’s a minimal system based on debian squeeze. I installed it over opensuse 11.4. I tried 2 versions of opensuse 11.4: Gnome and a network install of XFCE.

When I installed suse with XFCE I did just the bare minimum that I could. Started with a base system and left out what I didn’t need (bluetooth, accesability and disability features, firewall, apparmor, etc…) I also went into the services and disabled what I didn’t need.

Even at a minimum system it was pretty slow. Is it just the kernel OpenSuSE was using or are there other items it installed that I don’t know about that bog the system down?

I remember in the olden days I had to do an hdparm -c 1 -d 1 -k 1 /dev/hda command to try and speed up the read/write of the hard drive. Could crunchbang be having extra startup commands that suse doesn’t?

While I have crunchbang installed, what information should I gather when I migrate back to openSuSE so I can try and match the speed?

-edit-
I should mention that even though I love the speed of crunchbang, the OpenSuSE support community is the best I’ve seen, so I’m willing to sacrifice the speed for ease of use and support.

On 03/04/2011 09:06 PM, DupermanDave wrote:

> Even at a minimum system it was pretty slow. Is it just the kernel
> OpenSuSE was using or are there other items it installed that I don’t
> know about that bog the system down?

crunchbang does not use the same window manager or desktop
environment…both of which (in openSUSE) is much larger in code base
size and complexity, and therefore slower in speed…

> While I have crunchbang installed, what information should I gather
> when I migrate back to openSuSE so I can try and match the speed?

that is kinda like asking:

I’m now driving a nice Ferrari, when i move to the family station
wagon what should i bring with me so i can corner and accelerate then
like i do now?

> -edit-
> I should mention that even though I love the speed of crunchbang, the
> OpenSuSE support community is the best I’ve seen, so I’m willing to
> sacrifice the speed for ease of use and support.

then why ask how to speed it up?

if you want the speed then stick with crunch, or one of the other
small, light distros made for speed…

but, if you want a rotating cubes, rounded corners and other bling you
have to pay for it in code size and cpu cycles…no way around it.


DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

speed and suse do not necessarily go together

suse is a bit on the bloated side , if the machine is a bit underpowered then it will be slow

an example
Opensuse 11.3 64 bit on a box with a i5 cpu and 8 gig ram boots WAY SLOWER and i do mean WAY WAY SLOWER
than Arch and SL6 ( scientific linux 6 beta) on a 11 year old computer that has a 11 year old p4 cpu and 1 gig ram

Arch is almost a “no time to boot” about 15 sec from power off to running
and SL6 on that old box is about 30

suse 11.3( gnome) however is about 1.5 min. from off to running

On 2011-03-05 06:06, JohnVV wrote:

> suse 11.3( gnome) however is about 1.5 min. from off to running

Booting time is different from speed at running apps.

I can not comment on crunchbang as I don’t know what it is.


Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

On 03/05/2011 02:33 PM, Carlos E. R. wrote:
> I can not comment on crunchbang as I don’t know what it is.

its a skinny Debian based distro built for speed…which comes without
a lot of the things openSUSE desktop users want…

http://distrowatch.com/table.php?distribution=crunchbang


DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

I was still under the impression that “Linux is Linux.” When I first started using linux (redhat 5.0/6.0) About every distro was the same. Up until Ubuntu came out, the only real thing that separated the distros apart was that some came with tools like Yast and the Mandrake Control Center and things like that. Mandrake and SuSe automounted drives and CDs wheras the rest made you manually do the mounting. But underneath, linux was still linux and they were all about the same in speed.

Fast forward to now and it seems people have their own customizations, be it certain kernel versions or modifications to the kernel, or OpenSuSE with the apparmor and firewall. Now that I think about it, I guess it was a vague and stupid question, since I’m stuck in the “linux is linux” mindset when that’s not entirely true anymore. There’s tons of different package managers (whereas when I started there was Apt, Rpm, and slackware tar files.) And with each one some of done some kernel tweaking. I get it.

On 03/05/2011 06:06 PM, DupermanDave wrote:
>
> I get it.

no such thing as a stupid question…maybe just not completely
thought through…

of course, linux IS linux as long as you know when you say ‘linux’ you
really mean “linux kernel”…after that there are LOTS of variations
on file structure, configs, default services, on and on and on…

further, while the kernel in one distro might be similar enough to the
kernel in another distro, everything else can be so different as to
make huge differences in the speed you see (or not)…

speed is especially impacted by the amount of stuff going on in the
display…consequently there is a huge difference in apparent speed
between XFCE and KDE…but, if you want wobbly windows, rounded
corners, see though panes etc etc etc all of that stuff has to be
‘computed’ and moved to the display…


DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

My impression (and I haven’t done any back-to-back testing for over five years) is that if you put two Linux systems in the same configurations, the speed is essentially the same. That said, the defaults may be wildly different and one system may offer options that the other either doesn’t offer at all, or doesn’t easily offer.

While services are the thing that most people comment on, I think that set up for the disk partitions is something that is often overlooked. Firstly what are the respective partition types (probably ext3 or ext4, but reiser3 was fast, back in the day for general purpose usage, and not many people have been adventurous enough to go with reiser4)? Also the journalling options (assuming that you are using a journalling filesystem); Suse tends to be fairly conservative (low risk of data loss, rather than prioritising absolute speed).

Note also that side-by-side comparisons are difficult as partitions early in the disk will be faster than partitions late in the disk. Also, the partition layout (separate partition for home, ‘distance’ (stroke length) to swap, lots of partitions vs few partitions and how they are organised) makes a difference. So, you really can’t do side-by-side benchmarking, you have to do sequential installations, if you want valid timings.

Peculiarly, reducing swappiness is a popular ‘speed up tweak’ with Ubuntu users, but isn’t really popular anywhere else. (And, anyway, it is probably best described as increasing ‘responsiveness’ rather than ‘speed’, even when it works, and there is a suggestion that in some use cases it would be the opposite of constructive.)

And then there is GUI. How much of an impact this makes depends on how much RAM you have. The heavier weight GUIs use more ram (doh!), but if you have lots, that doesn’t really make much of a difference, but if you don’t have enough, then it could be a massive factor (…but then, on the other hand, using apps/applets from one GUI under another implies that you load the libs from both GUIs and that could be a ‘worst of both worlds’ situation, if you don’t have enough ram for that…).

As an example, which isn’t totally relevant, my preference is for KDE, but I’m finding that after a while it slows down, as it gets deeper into swap, so I’m currently trying out Gnome (which I don’t really like, but this is an experiment). I haven’t tried this out for long enough, but I’m not (yet??) seeing the slow down that I see with kde as its memory usage gradually rises and I’m not seeing the climb in CPU temperature. This seems potentially down to kde4 (or the apps?) leaking memory, and I wouldn’t really advise using KDE4.x for perf critical stuff, and there are probably big differences between different kde 4.x versions, so, unless the two distros use exactly the same kde 4 versions, it would be hard to compare.