Can Ubuntu 9.10 Outperform Mac OS X 10.6?

[BlackStar]

All that is is a measure of how many services are being loaded on start up, not how well the OS performs. Ubuntu loads a lot more than an install of ArchLinux, but then is capable of a lot more out of the box, e.g. I believe off the top of my head that by default ArchLinux doesn't install or load cups. If you want to print you've got to install it. Ubuntu does, and it runs on start up.

Actually, I ran some tests on this a few months ago and they indicated a strong correlation between disk accesses and boot times. I can't find the thread right now (it's on these forums), but the results went somewhat like this (fastest to slowest):

1. Arch: ~100MB
2. Ubuntu: 140MB
3. Fedora: ~200MB
4. OpenSUSE: ~250MB
5. Vista: ~350MB

With the exception of Vista, boot times where roughly linear with disk reads. Tests were carried out in brand new, fully updated VirtualBox VMs and the results were read by the VBox performance counters.

[Kano]

Bootperformance is not all, also it is like compareing apples and oranges. For example U does not mount any other partition thats not used by default - that means only / and maybe /home. My distro is configured to mount every partition because it is usally more usefull, especially when you run services like samba, nfs that provide those externally. On a system with really lots of partitions that must be slower. Also when you install extra services live vdr or so that will not speed up boottime, so the end result can be twice or more time compared to a pure install with 1 partition on a hd. Of course the extra fast times you only get with help of a SSD which is too expensive or too small for lots of people. Even without any special tricks a pure install boots in about 15-16s to kdm (using bootchart) on a 3 ghz dual or similar pc with 1 tb hd. If you disable the initrd and splash you can gain 2 extra seconds. But the time you wait till the bios is ready and grub shows is not measured in bootchart, combined it could be 1 min too - especially with lots of hds in a pc. Thats not that critical until you power off the system every few minutes. For laptops it could make a difference, there it would be also usefull that suspend works, thats possible with desktop pcs too, but usally not used.

[Garp]

btw. Ubuntu is a VERY slow linux-dist!!!

On my old Pentium 4(2GHz) with 512MB Ram, takes Ubuntu 40 sec to boot and is pretty slow agienst Windows XP. If I run Archlinux isted of Ubuntu, my old computer boots on 7 sec and the hole operating-system is twice faster then Windows XP!!

All that is is a measure of how many services are being loaded on start up, not how well the OS performs. Ubuntu loads a lot more than an install of ArchLinux, but then is capable of a lot more out of the box, e.g. I believe off the top of my head that by default ArchLinux doesn't install or load cups. If you want to print you've got to install it. Ubuntu does, and it runs on start up.

Real performance is what happens after the system has finished booting up and you actually start to use it.

If you want to do a performance comparison of Ubuntu and ArchLinux, install the Phoronix Test Suite in each and run it through the gamut of tests. Be sure to keep as much identical between them as you can, e.g. format the file system using the same format (e.g. ext3).
Then come back here, post your results, and then you might actually be speaking from a position of real knowledge about ArchLinux's performance over Ubuntu.

My Core2Duo workstation when it was running archlinux and doing everything I needed it to do took over a minute to boot even using sreader, up from ~20 seconds that it took to boot in it's clean install state because I was running a fair number of additional services like Apache, MySQL, Exim and so on. Heck I even had to install a bunch of services that Ubuntu installs by default.

[BlackStar]

If I run Archlinux isted of Ubuntu, my old computer boots on 7 sec

Bootchart or it didn't happen. 7'' boot-to-desktop is pretty much impossible without a solid-state drive - and the Arch forums confirm this.

I'll have to call shenanigans, especially in light of the "twice faster than XP!!" comment.

For comparison, my laptop boots Ubuntu 9.10 in ~7'' (~5'' to GDM + 2'' for desktop) and it's a much more beefy system than yours (Intel X25-M SSD, 2.5GB RAM, Core 2 @1.8GHz). Ubuntu is also much more responsive than Windows 7 in day-to-day use on this machine.

[BlackStar]

Wow, I coudln't imagine Linux was going so well in Greece.
So in the library installed few SuSE systems only? The current PCs that they are still used what run?

Oops, missed that question, so let me answer it now that the thread has been necroed.

The central PC labs now use Ubuntu 8.04 exclusively. I think a few PCs might be dual-booting to WinXP (for CAD/CAM software), but everything defaults to Linux now, which is pretty nice!

[leetsweden]

btw. Ubuntu is a VERY slow linux-dist!!!

On my old Pentium 4(2GHz) with 512MB Ram, takes Ubuntu 40 sec to boot and is pretty slow agienst Windows XP. If I run Archlinux isted of Ubuntu, my old computer boots on 7 sec and the hole operating-system is twice faster then Windows XP!!

Im OSX and Linux user, and I can say that Ubuntu (Gnome) isn't close to OSX in GUI, Windows is closer.

[Apopas]

Also Red Hat when it comes to share of the software industry doesn't even come close to ranking as a large software company.

That exactly was the positive thing I pointed in my post mate.

[L33F3R]

Microsoft funding the development on the Linux kernel? I think not! Microsoft is scared to death of Linux. Just look at itsbetterwithwindows.com.

lol i have xp installed on my fathers netbook and its slower then a turtle on the highway.

[deanjo]

How can you possibly know where and how all those people are employed?

Hang around the lkml and irc development rooms or go to pretty much any linux summit.

I don't know. You insinuate that many of them do, but provide no evidence. Google sells services, not software anyway, and runs linux based servers.. I fail to see the relevance.

Which is generated off of their closed source, highly protected search engine.

I didn't say there wasn't. Those are products that they sell, they aren't looking to anyone else for money.

MS has puchased a large amount of licenses from Novell. That capitol all filters down to the projects that Novell sponsors and as well as their freeing of patents.

[krazy]

A large number of independent developers working on the kernel work as well on proprietary software to pay the bills.

How can you possibly know where and how all those people are employed?

Also how many of those developers have received scholarships from organizations such as google, microsoft, etc for r&d and scholarships then use those techniques in the kernel or other opensource projects?

I don't know. You insinuate that many of them do, but provide no evidence. Google sells services, not software anyway, and runs linux based servers.. I fail to see the relevance.

as far as handouts go you might want to see who Novells largest customer has been the last couple of years and check their portfolio of solutions. You will find many proprietary solutions in there catalog.

I didn't say there wasn't. Those are products that they sell, they aren't looking to anyone else for money.