Speed Comparison between 23 distros

Hello fellow NZLinuxers Here a post of the results of the comparison between 23 distros on speed and many more categories. http://www.yoper.com/forum2/index.php?showtopic=953 Here a little ranking http://www.yoper.com/comparison/desktopcomparison.html Hope you'll enjoy it. Cheers Andreas

Here a post of the results of the comparison between 23 distros on speed and many more categories.
Do you have a description of your test methodology? Daniel

Hi Andreas, I realise that you of course are full of yoper propaganda, and rightly so given your position with respect to that distro :-) That aside, I too would like to know the methodoly behind said tests. IE: Things like buffer and cache reads, and possibly even user memory measurements, are well known as theoretical benchmarks, and, while they provide interesting trivia, quite often have no relevance to the real world. Stuff like 'How many megabytes per second can I pull off of a yoper based NFS server', or 'how many webpages per second does this fedora-based version of apache serve', stuff like that, are 'real world' examples and are more useful and will be much more accepted. While I'm not saying there's anything wrong with the figures that were provided, I feel they may create more controversy than good (ie: will leave people arguing pointlessly about yoper, etc, etc rather than winning anyone over to your cause). Also, while boot time *is* a useful and valid statistic, the question must be asked - booting what? I'm pretty sure I could write a custom mini-distribution with a kernel, bash and nothing else which would boot faster than any of the ones listed :-) Contrived example I know, but it illustrates my point. Sure the benchmark was done by a member of your community and hence you had no control over it, but reformatting it and posting it in the URL it was on leads people to believe that yoper ltd/incorporated/TM/whatever is the source of these benchmarks and hence you yourself will suffer any criticism of them. Be wary, the religious distro wars are afoot :-) Andreas Girardet wrote:
Hello fellow NZLinuxers
Here a post of the results of the comparison between 23 distros on speed and many more categories.
http://www.yoper.com/forum2/index.php?showtopic=953
Here a little ranking
http://www.yoper.com/comparison/desktopcomparison.html
Hope you'll enjoy it.
Cheers
Andreas
_______________________________________________ wlug mailing list | wlug(a)list.waikato.ac.nz Unsubscribe: http://list.waikato.ac.nz/mailman/listinfo/wlug

Orion Edwards wrote:
Sure the benchmark was done by a member of your community and hence you had no control over it, but reformatting it and posting it in the URL it was on leads people to believe that yoper ltd/incorporated/TM/whatever is the source of these benchmarks and hence you yourself will suffer any criticism of them. Be wary, the religious distro wars are afoot :-)
OK never mind it was not written by any old member of the community but by the "Yoper business development coordinator"

If you want a really relevant test for a desktop distro, try how fast did the unit boot , then from that fresh boot, how long to load open office. and with one of the big spreadsheet tables with lots of math. how long was the recalc time. even though I make hardware for a living, and live with these sorts of measurements all the time, if also find them meaningless in the real world. but even more sadly, my 2 distros that i use are not compared, but as one runs faster than a normal slackware install, its better than 7 on the list,but since the test could only be carried out on the hardware used for the original, I cannot compare them even if I did the same tests. .. pity. Andreas Girardet wrote:
criticism of them. Be wary, the religious distro wars are afoot :-)
OK never mind it was not written by any old member of the community but by the "Yoper business development coordinator"
And since when is business a bad thing? Are you a communist LOL :)
Andreas
_______________________________________________ wlug mailing list | wlug(a)list.waikato.ac.nz Unsubscribe: http://list.waikato.ac.nz/mailman/listinfo/wlug

runs faster than a normal slackware install, its better than 7 on the list,but since the test could only be carried out on the hardware used for the original, I cannot compare them even if I did the same tests. .. pity.
You are beine ridiculous. There are 300+ distros in the world! No one can test all distros! Andreas

Andreas Girardet wrote:
runs faster than a normal slackware install, its better than 7 on the list,but since the test could only be carried out on the hardware used for the original, I cannot compare them even if I did the same tests. .. pity.
You are beine ridiculous. There are 300+ distros in the world! No one can test all distros!
I found it odd that if you factored in "third party support", the Yoper-friendly list was almost turned 100% upside down. Craig

Actually I only meant to comment on the complexity of the tests. I did not mean to say anything else. I think most users will just look at the table and put it in the too hard basket. It was an observation, and not a complaint. Andreas Girardet wrote:
runs faster than a normal slackware install, its better than 7 on the list,but since the test could only be carried out on the hardware used for the original, I cannot compare them even if I did the same tests. .. pity.
You are beine ridiculous. There are 300+ distros in the world! No one can test all distros!
Andreas
_______________________________________________ wlug mailing list | wlug(a)list.waikato.ac.nz Unsubscribe: http://list.waikato.ac.nz/mailman/listinfo/wlug

Gavin Denby wrote:
If you want a really relevant test for a desktop distro, try how fast did the unit boot , then from that fresh boot, how long to load open office.
and with one of the big spreadsheet tables with lots of math. how long was the recalc time.
No a bad idea actually. Perhaps WLUG should organize some more useful benchmarks of the different distros using typical desktop applications.

While said in jest, it is sadly the image most linux users get tagged with, and the promotion of tandoori nachos or whatever his name is in wellington, did nothing to help with the linux image. Thank heaven for linspire and IBM putting on a good image.
And since when is business a bad thing? Are you a communist LOL :)
Andreas

* Gavin Denby <redhat(a)ihug.co.nz> [2004-07-11 23:28]:
Thank heaven for linspire and IBM putting on a good image.
Linspire give me an awfully bad aftertaste, not the least because they're charging money for services to which they add no value and which users of their distro can get gratis. By default the distro also logs the user in as root -- without even telling the user that something different is possible, let alone desirable. Or at least these were the facts back when I was marginally interested in their products, before I found out about them and lost any interest whatsoever. I will certainly agree that IBM is doing a lot for Linux. The other name that should be cited is not Linspire, but Novell. The funny thing is, every single major player save for Microsoft has decided that they can build a business on libre software. Microsoft is the only company among the top players in the IT field that are railing against freedom. We'll have to see how long it takes for the general public to take note.. Regards, -- Aristotle "If you can't laugh at yourself, you don't take life seriously enough."

The funny thing is, every single major player save for Microsoft has decided that they can build a business on libre software. Microsoft is the only company among the top players in the IT field that are railing against freedom. We'll have to see how long it takes for the general public to take note..
On another note, it would seem that every single major player, save for Red Hat, has decided that they can build a business on libre software but by adding value to it, and keeping that value commercial/proprietary. Up until recently you couldn't download SuSE isos (I don't know if the ones you can get now are up to date and the same as the ones they sell), and all the Netware stuff on Linux won't be free. Microsoft don't need to do this because they already have the base of another operating system. Craig

* Craig Box <craig(a)dubculture.co.nz> [2004-07-12 00:24]:
On another note, it would seem that every single major player, save for Red Hat, has decided that they can build a business on libre software but by adding value to it, and keeping that value commercial/proprietary.
Partially. The stuff that's interesting to businesses remains commercial, but IBM and Novell are known sponsors of a *lot* of libre software development.
Up until recently you couldn't download SuSE isos (I don't know if the ones you can get now are up to date and the same as the ones they sell), and all the Netware stuff on Linux won't be free.
Yeah, and once they were taken over by Novell whose business model is more solid than selling CDs to end users it was opened up. Regards, -- Aristotle "If you can't laugh at yourself, you don't take life seriously enough."

Linspire is not my favourite either, but they have made a simple and valid commercial model which is right now the only linux you buy preinstalled, not ideal in any stretch, and yes they support wine and a few other os projects too. and they click and run has made a lot of friends. the fact they are making money and enemies, is helping with IMAGE , Yes i agree on novell, Suse and Ximian makes a great base to build from, and IBM have given heaps, but mandrake and lindows are the commercial ones that get the scrutiny in ny neck of the woods, and so far its good news. BTW the Munich tender went suse and the Paris City contract went Mandrake, so the IMAGE in the commercial world is getting a lot more attention. Nowdays when I tell a client we use an embedded linux, they just ask about support and do their people need to know anything about config, once I show them a web interface, they aren't interested in the os. so things are better than they were 3 years ago.
Linspire give me an awfully bad aftertaste, not the least because they're charging money for services to which they add no value and which users of their distro can get gratis.
By default the distro also logs the user in as root -- without even telling the user that something different is possible, let alone desirable.

For any that didn't know. and in light of questions about who give what back, I found this release informative. but also raised questions. Linspire (lindows) have released PhoneGaim for Lindows, source for other OS's and are planning a windows release soon. PhoneGaim is an extension of the gaim project, which Linspire support financially and will be added to the gaim project at their release cycle. phonegaim adds sip features into gaim, and allows pc-pc (free) and pc-phone calls with apropriate accounts via a sip network. More info at www.phonegaim.com OK enough media, now question time We are planning to include this in our (CollegeLinux) 2.6 release, but wondered if anyone new of a sip server that linux users can use with sip features, having a local sip network could be very helpfull in some scenarios, and could effectively make a local PABX. we are already planning a jabber network and jabber extensions on gaim so that users can form their own private IM platform, but what about voice IM too. does anyone know of such a server ? otherwise I'll keep serching.
Linspire give me an awfully bad aftertaste, not the least because they're charging money for services to which they add no value and which users of their distro can get gratis.

Gavin Denby wrote:
We are planning to include this in our (CollegeLinux) 2.6 release, but wondered if anyone new of a sip server that linux users can use with sip features, having a local sip network could be very helpfull in some scenarios, and could effectively make a local PABX. we are already planning a jabber network and jabber extensions on gaim so that users can form their own private IM platform, but what about voice IM too.
does anyone know of such a server ? otherwise I'll keep serching.
Sounds like you want Asterisk (or *) Cheers Philip Murray

Yep its the one were are looking at, it will not be on the CD, but we will add it to the download repository. Just wondered if anyone was using it, and what they thought of it since we use gaim anyway, and the phone addon is a worthwhile feature, it seemed like an idea to have a server to play with too. Its the sort of thing someone doing a masters in business, should be exposed to .... IMHO Philip Murray wrote:
Gavin Denby wrote:
We are planning to include this in our (CollegeLinux) 2.6 release, but wondered if anyone new of a sip server that linux users can use with sip features, having a local sip network could be very helpfull in some scenarios, and could effectively make a local PABX. we are already planning a jabber network and jabber extensions on gaim so that users can form their own private IM platform, but what about voice IM too.
does anyone know of such a server ? otherwise I'll keep serching.
Sounds like you want Asterisk (or *)
Cheers
Philip Murray
_______________________________________________ wlug mailing list | wlug(a)list.waikato.ac.nz Unsubscribe: http://list.waikato.ac.nz/mailman/listinfo/wlug

Yep its the one were are looking at, it will not be on the CD, but we will add it to the download repository.
Just wondered if anyone was using it, and what they thought of it
Yes, im using asterisk, and have wikiid some info on it..
since we use gaim anyway, and the phone addon is a worthwhile feature, it seemed like an idea to have a server to play with too. Its the sort of thing someone doing a masters in business, should be exposed to .... IMHO
Ive been using asterisk at home to connect all my sip devices (laptops with software phones and a hardware phone) to various providers, sipphone.com, freeworlddialup etc. Its a very cool system, you need lots of bandwidth though (which luckily i have) and it can make toll calls very cheap even to PSTN if you are prepared to route your calls from NZ to USA and back to NZ ;-)

This little poem's been getting a lot of airplay lately. Apparently it's being aired by Total Furniture, but the first time I heard it I was honestly expecting to find that Microsoft was behind it.. "It is unwise to pay too much, but it's worse to pay too little. When you pay too much, you lose a little money - that is all. When you pay too little, you sometimes lose everything, because the thing you bought was incapable of doing the thing it was bought to do. The common law of business balance prohibits paying a little and getting a lot - it can't be done. If you deal with the lowest bidder, it is well to add something for the risk you run. And if you do that, you will have enough to pay for something better." John Ruskin 1819-1900

Hi Andreas, I realise that you of course are full of yoper propaganda, and rightly so given your position with respect to that distro :-) That aside, I too would like to know the methodoly behind said tests.
The person who did the test is very very good, that is all I can say. If you want his email, contact me offlist.
IE: Things like buffer and cache reads, and possibly even user memory measurements, are well known as theoretical benchmarks, and, while they provide interesting trivia, quite often have no relevance to the real world. Stuff like 'How many megabytes per second can I pull off of a yoper based NFS server', or 'how many webpages per second does this fedora-based version of apache serve', stuff like that, are 'real world' examples and are more useful and will be much more accepted.
Good idea and I am sure that stuff like that would look good in Yoper too, since I have added tweaks to the TCP stack into the startup that no other distro has.
While I'm not saying there's anything wrong with the figures that were provided, I feel they may create more controversy than good (ie: will leave people arguing pointlessly about yoper, etc, etc rather than winning anyone over to your cause).
what can I say .... since when is proof so decremental to winning people over?
Also, while boot time *is* a useful and valid statistic, the question must be asked - booting what? I'm pretty sure I could write a custom mini-distribution with a kernel, bash and nothing else which would boot faster than any of the ones listed :-) Contrived example I know, but it illustrates my point.
Why don't you do it then. Every one has a choice to create their own distro. Every distro has a choice of doing what they are doing and my choice was to make it fast. Another user of mine has writtens scripts that boot in 20 seconds. I haven't added them since I agree with you that it is at some point fairly pointless and I rather keep them standard for users to at least know about them. I also am going to integrate Yast2, which needs some sort of standard core. Fact remains though that performance, which is the purpose of Yoper, since I personally like that :) ..... is an important aspect of Yoper. the init scripts are lifted from RedHat and as such it should show one thing, that the same scripts booting in 2 different OS's can certainly be faster in one than in the other.
Sure the benchmark was done by a member of your community and hence you had no control over it, but reformatting it and posting it in the URL it was on leads people to believe that yoper ltd/incorporated/TM/whatever is the source of these benchmarks and hence you yourself will suffer any criticism of them. Be wary, the religious distro wars are afoot :-)
Personally I like all distro's and that is not just something I say willy nilly. I work (or shall I say advocate) daily on the job while administer with various distros and promote various distros and do not really give a darn about which one others use. I am not a zealot who thinks that there can only be one. The strenght of Linux is in fact that there are many many distros for various purposes and Yoper is just one of them. The purpose of Yoper is to be very fast and to have the nicest features of other distros. Why? Because I am a learner and I learn about different distro's by integrating things they have in my own distro. Why have I chosen to do my own? Because I was sick and tired of rattling hard disks, just because I load an application. I like it that way. Why have I reformatted the findings? After spending so much time doing Yoper and even more time implementing RedHat, SuSE, Debian or whatever other distro I work with I wanted to know black and white that what I have done so far was actually the right direction of work. All in all if this just keeps people off using Yoper, so be it, even though I find it quite odd, especially from a Kiwi perspective a little bit support would be great, after all it is a Kiwi distro. I also find it odd that such a comparison (any comparison is byist) should stop people from using it? Why would that be. Personally I like doing Yoper because it offers me a chance to deepen my skills that I use professionally. If people use it great, if they don't so be it. Anyhow ... I hope this explains a little more. I did not want to stir you up, just though you might be interested. Cheers Andreas

My point was not to try belittle yoper or the benchmarks, but I was merely stating from experience that posting a lot of theoretical results to a community which has a tendency to get fired up about things, can often result in a lot of fighting and put-downs of said benchmarks and a loss of face for the poster thereof. My experiences were from the gaming/overclocking type community so I've seen my fair share of benchmarks and backlashes and zealotry about them. The linux community, although it likes to think of itself as far far above the pitiful ignorant gamer, is not all so different, at least when it comes to bull headed zealotry. Remember all those 'Microsoft is cheaper and better than linux' benchmark type things? The end result of those (at least amongst everyone I've talked to about it) was that said company was full of lies and disreputable - those 'independant' research companies all lost face and microsoft remained the big evil troll it always was. All it would take for the same thing to happen to those yoper benchmarks is some equally fanatic redhat supporter to post some other benchmarks, perhaps of 'real world' tests to debunk them, and everyone would go around saying how yoper sucked because it only posted the benchmarks it won in and ignored the others... whether this is true is beside the point - zealots will always be zealots. So yeah, not attempting to put you down at all, but just giving what I'd like to think is a word of warning. People will naturally be very skeptical about a benchmark done by yoper people and posted on the yoper site.
Why have I reformatted the findings? After spending so much time doing Yoper and even more time implementing RedHat, SuSE, Debian or whatever other distro I work with I wanted to know black and white that what I have done so far was actually the right direction of work.
All in all if this just keeps people off using Yoper, so be it, even though I find it quite odd, especially from a Kiwi perspective a little bit support would be great, after all it is a Kiwi distro. I also find it odd that such a comparison (any comparison is byist) should stop people from using it? Why would that be.

So yeah, not attempting to put you down at all, but just giving what I'd like to think is a word of warning. People will naturally be very skeptical about a benchmark done by yoper people and posted on the yoper site.
Not a problem and I can understand, what you are saying. Personally I rather concentrate to get people using Yoper with no or little Linux experience than people who are already convinced of their own distro. I think that the first lot of people might be a little bit more open about reading such a comparison and giving Yoper a try. Skepticism is good and I certainly appreciate that since it just shows what a smart bunch of people hard core Linuxers are. If we would have put it on another website ... it would not have changed the fact that it was done by a very engaged Yoper user who also happens to want to build a business up in the US around Yoper. He was a user first and now he has business interest, which is why I made him business coordinator, since I am a geek and that is my role in the project :) ..... The final Version2 release will be announced with the next day ..... so if any waikato kiwi out there wants to give it a go, regardless of personal choice and give me some positive critizism, then I am will be very glad to build that into the next releases. Cheers Andreas
participants (9)
-
A. Pagaltzis
-
Andreas Girardet
-
Craig Box
-
Daniel Lawson
-
Gavin Denby
-
Kyle Carter
-
Orion Edwards
-
Philip Murray
-
zcat