Is Arch Linux Really Faster Than Ubuntu?

[V!NCENT]

Agreed on Arch being faster on my older system, but remember that 32-bit Arch Linux is i686-optimized, while 32-bit Ubuntu has not been up to this point (that's about to change). Note that the article uses 64-bit installs.

I think this is the problem everbody is dealing with here; load.

Modern computers have multi-core CPU's running at 2-3 gHz, RAM space counting in the gigabytes. Modern graphics cards measure their insane parallelisation in the theraflops.

These modern systems can handle load easiliy. In the case of modern computers you can achieve greater speed under a greater load. With less load comes less speed.

But what if you're having an older system? Parallel would kill you. The background services, drivers, applications, the GUI all run on the smae thread (with time slicing) as your favorite game. This is slowing things down horribly. So the problem is not speed on these systems, it is just load.

This ammount of load to increase speed in devastating for older computers and they crumble onder it.

So Ubuntu is faster on modern systems where Arch is actually slower.
On older systems Arch is faster than Ubuntu.

Load != Speed

[BlackStar]

Well, Ubuntu runs smoothly on my Atom netbook (1GB RAM, slow disk, dog-slow CPU, non-existend Intel GPU), so I'm not sure that comparison makes sense.

Arch will likely perform better on low-memory environments (<512MB) but those are not common nowadays. Frankly, it's not as if a modern Arch desktop (gnome, gnome-extras, compiz, openoffice, firefox or their KDE equivalents) is all that different from the Ubuntu desktop in memory usage, disk requirements or compiler optimizations. The main difference is that Ubuntu/Kubuntu ships a different theme, a couple of different apps and a few different patched by default - big effing deal!

(You can compare them on VirtualBox if you don't believe me. The difference is pretty small and Ubuntu boots faster, too.)

[HoboJ]

This comparison confirms everything I've known for a while. My arch install isn't any faster than my ubuntu install. However, my arch install is consideraby more snappy than my ubuntu install because it's heavily customized to be light and quick.

[RealNC]

Windows 7 pwns them both, so there.

[BlackStar]

Windows 7 pwns them both, so there.

Until you try to use it on VirtualBox. It makes for an, err, interesting experience.

[ethana2]

So you want to replace the Windows monoculture with an Ubuntu one. Thanks, but no thanks. Besides I don' think that would be sustainable. The Linux ecosystem relies heavily on the contributions from companies such as Red Hat. Not so much Canonical unfortunately.

Yep. ..I purchase software in the Android Market all the time, I look forward to being able to do the same in the Ubuntu Software Center. I also look forward to having a fixed monthly free software donation rate in an ubuntu one account that gets distributed to my favorite ubuntu packages according to my usage.

Having multiple target platforms is a pain for developers. The reason I don't use Windows isn't because it's too common, it's because it's not good enough. I look forward to having a Free Software based drop in replacement.

[curaga]

Well, Canonical does contribute more than Microsoft, so Ubuntu having 90% marketshare wouldn't be as bad.

[kbios]

Sorry, can you post evidence of that? Just tried Unigine Heaven under kde 4.4.3:
With effects manually deactivated reported FPS is more than 40 and the animation is smooth.
If I start it (fullscreen of course) with effects activated I get around 30 fps, but the animation is not smooth at all, as if it was doing 20 fps at most.

Never mind, I've found it. It's well hidden: you have to put UnredirectFullscreen=true under [Compositing] section in ~/.kde/share/config/kwinrc and restart kwin.

[yoshi314]

Blah blah, the same old drivel: "it feels faster to me", "yeah, the benchmarks aren't valid because I don't do that", "Ubuntu sucks because Compiz works as it should (wait what?)"

The cold, hard numbers show Ubuntu performing identically to Arch. Previous benchmarks show it performing on par with Gentoo. Get over it.

those benchmarks test how the system performs under some heavy task. but not the general user experience when doing regular work.

install ubuntu and arch on an older system (something from 2004 would be good, that's what i have at work atm) or an older laptop. and try doing some normal work like writing text, web, running a few heavier apps at once, git/svn/mercurial work. and run a webserver with cgit on top of that from that old pc.

i can tell you straight, arch will work better. i know because i've been using it at work for the last 3 years, even though i've been using gentoo since 2004 on my home pc.

package manager is very fast, system is very responsive. and most importanly the package manager doesn't try to outsmart me and run or configure system services for me. that's actually very helpful.

[V!NCENT]

those benchmarks test how the system performs under some heavy task. but not the general user experience when doing regular work.

In other words: load the CPU to the max and the speed is the same. Having to timeslice through the usual per time versus timeslicing throught the usual minus backgroudn services that check periodically and other services enables by default that one does not use means more CPU time for you favorite DE/WM and thus: snappier.