Benchmark: Ballnux is Fat

Dr. Roy Schestowitz

2008-12-15 11:08:09 UTC
Modified: 2008-12-15 11:08:09 UTC

Stevie and SUEsie

Phoronix tests show that SUSE is the fattest among all.

What we gathered from these tests conducted on the Intel Atom desktop was that Ubuntu 8.10 was generally the fastest distribution. OpenSuSE 11.1 RC1 on the other hand was in last place most frequently.

Why would anyone choose SLES/SLED 11 over the competition, which is likely to be lighter and more agile? ⬆

Comments

AlexH

2008-12-15 11:22:27

Nice leap from "unreleased/beta version of OpenSUSE" to "SLES/SLED", Roy...
Jo Shields

2008-12-15 11:25:00

Wasn't he complaining because the fastest supercomputers in the world ran SLED, as per top500?
Roy Schestowitz

2008-12-15 11:30:41

The fastest one runs Red Hat.
AlexH

2008-12-15 11:35:06

Which is everything to do with the OS and nothing to do with the hardware, natch :D
Jo Shields

2008-12-15 11:43:53

Not an enormous shock. IBM's xCAT management stack is all RH-based
Roy Schestowitz

2008-12-15 11:48:31

HPC is not the point anyway.
Jo Shields

2008-12-15 11:51:11

The point of what?
Roy Schestowitz

2008-12-15 11:54:17

This post. The point I was making is that a release that's just days away (I can't imagine it contains much debugging code) is slower than equivalent counterparts and it's the basis for Novell's open source crown jewel, SLE.
AlexH

2008-12-15 12:01:02

@Roy: "the release" isn't slower, a couple of benchmarks are: Sunflow and Sqlite. In all other tests, all the distros performed basically the same.

And at least the Sqlite result is explained by ext3 barriers, which is a feature, not a bug.
Roy Schestowitz

2008-12-15 12:05:20

They are anatomically similar. There is not the expectation to find much discernible difference, but where it does exist, OpenSUSE is behind.
AlexH

2008-12-15 12:09:28

@Roy: sorry, but nonsense.

The Sqlite test is slow for one reason: it does systematic INSERT without placing the script in an sql transaction. That will be slow, because Sqlite guarantees that each individual insert statement hits the disk.

Why is OpenSUSE slow? Because it uses a more modern sqlite and disk FS, and the "fast" distros are open to potential data loss. Is it actually slow? No, because any real application would put those INSERTs in a transaction.

I know Sqlite very well, and I've read the Phoronix test code. Your attempt to interpret these results for me won't wash, sorry. OpenSuse isn't slow here, it's actually doing things right.
Roy Schestowitz

2008-12-15 12:16:40

There are more benchmarks where OpenSUSE loses. Watch all the pages of the article.
AlexH

2008-12-15 12:19:20

I did read them all. Sunflow is the only other one it loses out in a big way.

And y'know, that could just be a bug, which could be fixed, like those other distros who've been released for a while and received much wider testing....
Roy Schestowitz

2008-12-15 12:52:41

The differences in other tests are still pretty considerable.
Jo Shields

2008-12-15 13:11:16

Even the ones where Fedora is worst?
AlexH

2008-12-15 13:16:55

@Roy: it's slightly slower in some test which require disk i/o, it's slightly faster in tests which use OpenGL.

I just think it's remarkable that you feel very comfortable drawing conclusions where the people actually running the tests say they won't. It would have been fairer to them to quote them properly, not just the bits you liked.
Roy Schestowitz

2008-12-15 13:25:46

Michael is always careful not to show bias. Let's wait and see what happens when OpenSUSE goes GM.
AlexH

2008-12-15 13:44:08

@Roy: you're still making the assumption that the benchmark shows distro problems.

It doesn't matter whether OpenSUSE is GM or not with regards to the Sqlite benchmark; it won't "improve" and I've shown you why.

And let's say disk i/o is slower because of the write barriers. That doesn't make OpenSUSE "fatter" either.

It's really unfortunate that you attempt to write off the good work of the OpenSUSE community like this. They're producing an up-to-date and interesting distribution, and even if you don't use it, Ubuntu and Fedora (and others) would be less advanced without it. Shame.
Roy Schestowitz

2008-12-15 13:48:23

It doesn’t matter whether OpenSUSE is GM or not with regards to the Sqlite benchmark; it won’t “improve” and I’ve shown you why.

This contradicts your first comment
AlexH

2008-12-15 13:52:21

Good grief. Really sad that I have to spell this out: the Sqlite benchmark is slow because Sqlite is waiting for the hard disk.

That has nothing to do with your claims about SLED/SLES which you obviously cannot backup.
Jo Shields

2008-12-15 13:54:00

How are the two comments even remotely related?
Roy Schestowitz

2008-12-15 13:56:51

AlexH,

It's nothing to do with Sqlite. It loses in several other tests too.

Face it. OpenSUSE (RC) is fat, other are fast. Unless you can come up with other tests, the fact remains.
AlexH

2008-12-15 14:02:29

Roy, you keep moving the goalposts. First it's my SQLite comment is 'contradictory', then "it's nothing to do with Sqlite".

You're losing the argument. You're drawing conclusions from some raw data, and the people running the test say explicitly that they're not willing to draw definitive conclusions.

I've shown you an example where the benchmark suite is showing a difference which you would expect, and is not a bug, and is not OpenSUSE being "fat".

OpenSUSE is faster and slower on other tests, and there's a disk i/o issue which is explained by the ext3 write barriers. Again, not "fat".

Your conclusion is faulty because your data don't support it. Your "facts" are your interpretation.
Roy Schestowitz

2008-12-15 14:06:24

No, the only reasoning Michael gives is the stage of the release (RC). He hardly ever draws any potentially discriminatory conclusions.
AlexH

2008-12-15 14:08:48

Ok, so we'll just ignore all those problems with the benchmark that I raised, because we must have results which fit our conclusion!

Thank goodness for the scientific method.
Dan O'Brian

2008-12-15 14:11:22

Hmm, perhaps Roy should read the article. Michael has noted that the openSUSE 11.1 might be slower in a bunch of the categories because of debugging options turned on due to it being a pre-release.
Roy Schestowitz

2008-12-15 14:12:46

PTS is no gold standard but a de facto one. I correspond with Michael sometimes, so I'll point him to your suggestions.
Roy Schestowitz

2008-12-15 14:13:58

Yes, I noted this earlier.
AlexH

2008-12-15 14:19:04

I can’t imagine it contains much debugging code

Interesting way of noting debugging options may be turned on :D

Look, the benchmark results are interesting and all, and although I'm sure the benchmark can be improved (and I'm willing to help Michael with that, tell him) what is it really saying? It's saying that stuff changes.

You can go through every distro and there will be issues here and there. Benchmarks are good for showing areas which may need improvement (or, at least, knowing how things are changing - e.g. write barriers), but they're not good indicators of the quality of a distro.
Roy Schestowitz

2008-12-15 14:21:02

These samples of work are the least one can use to compare.
Diamond Wakizashi

2008-12-15 16:11:43

Don't expect quality from Microsoft or it's nasty Novell subsidiary.
neighborlee

2008-12-15 19:09:47

Why anyone would use a distribution that knowingly protects 'only' its usersbase from a monopolistic , known linux hating company whom also dove into the gutter by pushing their 'power' behind ooXML, is obvious in that not enough people know about it and those that do are confused by those who would obfuscate the facts for their own ends .

Thankfully the greater open source community was built around ideals entrenched in fairness , equalty and a 'one for all' ideology which in the end will rescue its users from the proverbial grips of Microsoft and its legions who obviously have no interest in said goals .
Slated

2008-12-16 09:29:53

Watching all these Microsoft/Novell fanboys squeal is hilarious.

SUSE = FAIL.

Get over it and move on ... preferably to something better, and untainted by the Vole.
Jo Shields

2008-12-16 10:14:23

"Fanboy" is a convenient word, but one usually relegated to squealing brats arguing about their favourite Pokemon.

Are you really at that level? "Yuo are teh dum coz u like charmander lolololo"? No? The drop shit like "fanboy" now.
Slated

2008-12-17 18:23:21

'“Fanboy” is a convenient word'

Yes - the most convenient aspect is that in this case it happens to be true. If you want to describe yourself as a squealing brat, then go right ahead, but I simply meant to imply that you are blinded to the truth by your irrational loyalty to something harmful.

Prove me wrong by switching to a less-tainted development platform that doesn't support Microsoft's anti-Linux agenda.

As for your AOL-speak gibberish, you'll have to translate that for me.

"drop shit like “fanboy” now".

No.

You drop shit like .Net/Mono now.
jo Shields

2008-12-17 19:50:06

Did you REALLY just use "I know you are but what am I"?

Kids -_-
Roy Schestowitz

2008-12-17 19:57:46

At least he passed the profanity filter. No s*.
Slated

2008-12-19 03:43:20

'Did you REALLY just use “I know you are but what am I”?'

No, I challenged the prudence of you supporting Microsoft's anti-Linux agenda with their encumbered development framework, and you completely failed to respond to that challenge, presumably because you can't.
jo Shields

2008-12-19 07:22:46

No, I challenged the prudence of you supporting Microsoft’s anti-Linux agenda with their encumbered development framework, and you completely failed to respond to that challenge, presumably because you can’t.

OR, alternatively, I don't give a poopie what you think; don't see the world through the same twisted spectrum; and have no interest in going "ZOMG YUOR RITE!" and stopping my packaging work on a set of interesting Free Software projects.

Really, you think "You drop shit like .Net/Mono now." is going to elicit ANYTHING more than utter contempt?
Roy Schestowitz

2008-12-19 09:18:10

If it was asked more politely, would that make a difference?
neighborlee

2008-12-19 20:05:46

@Jo Really now,,your childish outbursts leave you with zero credibility, thus you have lost your entire arguement.

I appreciate that though , because you just show what a pathetic arguement machine the mono supporters are embracing ;)

cheers ;) nl
jo Shields

2008-12-19 20:24:05

If it was asked more politely, would that make a difference?

Sure. As I've said before, I've had productive and informative discussions with mature & adult Mono critics before. There are none on this site though. Especially not Slated or neighborlee or yourself.

Really now,,your childish outbursts leave you with zero credibility, thus you have lost your entire arguement.

Sorry, that's not how Godwin's Law works. And perhaps you don't understand what an "outburst" is. I've thus-far resisted any such thing - although perhaps not according to your broken behavioural compass, I'm pretty sure I'm well placed to say what is & isn't an outburst where I'm involved. I'm good at spotting them.

I appreciate that though , because you just show what a pathetic arguement machine the mono supporters are embracing ;)

Yeah, because that's really the point of this little sidetrack, honest. Yep. Yes siree. And the point of this (debunked) article, too. Yep, that's definitely it, it's "a pathetic arguement machine". Right, yeah, sure.

Imagine a world where I give even a fraction of a poopie what you think. Hold on to that thought, as it only exists in your mind.
Roy Schestowitz

2008-12-19 21:13:26

Who ever mentioned Godwin’s Law?
Slated

2008-12-20 22:09:33

@jo Shields

"I don’t give a poopie what you think"

Well you obviously do, otherwise you wouldn't keep responding.

'Really, you think “You drop shit like .Net/Mono now.” is going to elicit ANYTHING more than utter contempt?'

You mean like 'drop shit like “fanboy” now'?

A FOSS traitor *and* a hypocrite. It gets worse.
jo Shields

2008-12-20 23:18:36

Well you obviously do, otherwise you wouldn’t keep responding.

Yeah, you got me. As I lay there at night under my Microsoft bedsheets, cuddled up to my plushie Steve Ballmer, all I can think about is what some guy with an angry blog thinks about me.

It's tough to imagine life if you weren't my friend :'(

You mean like ‘drop shit like “fanboy” now’?

Don't use words if you don't understand what they mean. It's pretty clear you don't.

A FOSS traitor *and* a hypocrite. It gets worse.

Oh yeah, I love betrayin' things. It's like heroin to me.
Roy Schestowitz

2008-12-21 00:13:39

Let's discuss this without catfighting. We can reach conclusions together. Tomorrow I'll write about GIMP being threatened by a patent lawsuit.
jo Shields

2008-12-21 00:25:21

Tomorrow I’ll write about GIMP being threatened by a patent lawsuit.

It has? Shit, from whom?

That's news that is!

As for the original post: 1 known upstream bug (which affects only intel hardware and also affects other distros such as Ubuntu 9.04), and one feature; I don't see either as meaning an unreleased SLES (which isn't even being benchmarked here) is "fat", despite your suggestion.
Roy Schestowitz

2008-12-21 00:28:59

That’s news that is!

It's not news, but a case of C&D that was kept quiet, AFAIK.

BTW, I hope to convince you guys, not to confront.

I'll post about it maÃÂ±ana. I'm too tired and I've been out drinking for two nights in a row.
Slated

2008-12-21 04:00:28

@jo Shields

Your sarcasm notwithstanding, you are nonetheless betraying the principles of Free Software by supporting Microsoft's Intellectual Monopoly, and the fact that the only argument you have to defend that position is vacuous sarcasm indicates you have already accepted that position is fundamentally indefensible, but you think that peppering this thread with your juvenile commentary will somehow distract others from that fact, and dissuade me from speaking the truth.

Guess again.
MonkeeSage

2008-12-21 06:41:27

[quote]...you are nonetheless betraying the principles of Free Software by supporting Microsoft’s Intellectual [b]Mono[/b]poly...[/quote]

Then GNU / FSF are also betraying the principles of Free Software, since they have their own C#-to-CLI compiler, CLR, and CLI disassembler, just like the mono project (and you can target pnet from monodevelop, BTW).

But...that must be some kind of fluke...RMS hasn't really jumped into bed with the Evil Empire, has he?!

"Our main goal is to make it easy to write portable application programs which work well both on DotGNU Portable.NET [b]and on Microsoft's .NET platform[/b]."[1]

Oh, cruel fate! Nothing is sacred anymore! Now we all need to boycott GNU! Immediately uninstall coreutils, GCC, and all other GNU software! We can't support software released by a monopoly-supporting hater of freedom like the FSF! (My irony-meter has spontaneously combusted, and is smoldering in the corner).

____ [1] http://www.gnu.org/software/dotgnu/pnet.html
MonkeeSage

2008-12-21 07:51:24

Ps. I love how Novell is so tight with Microsoft they make viral marketing campaigns trashing Vista (http://youtube.com/watch?v=GVOnFdMf0RU -- "p-leather", heh).
MonkeeSage

2008-12-21 07:54:05

Oops...original comment got lost...repost:

...you are nonetheless betraying the principles of Free Software by supporting Microsoft’s Intellectual Monopoly...

Then GNU / FSF are also betraying the principles of Free Software, since they have their own C#-to-CLI compiler, CLR, and CLI disassembler, just like the mono project (and you can target pnet from monodevelop, BTW).

But...that must be some kind of fluke...RMS hasn't really jumped into bed with the Evil Empire, has he?!

Our main goal is to make it easy to write portable application programs which work well both on DotGNU Portable.NET and on Microsoft's .NET platform.[1]

Oh, cruel fate! Nothing is sacred anymore! Now we all need to boycott GNU! Immediately uninstall coreutils, GCC, and all other GNU software! We can't support software released by a monopoly-supporting hater of freedom like the FSF! (My irony-meter has spontaneously combusted, and is smoldering in the corner).

____ [1] http://www.gnu.org/software/dotgnu/pnet.html
MonkeeSage

2008-12-21 07:55:04

Oops...original comment got lost...repost:

"...you are nonetheless betraying the principles of Free Software by supporting Microsoft’s Intellectual Monopoly..."

Then GNU / FSF are also betraying the principles of Free Software, since they have their own C#-to-CLI compiler, CLR, and CLI disassembler, just like the mono project (and you can target pnet from monodevelop, BTW).

But...that must be some kind of fluke...RMS hasn't really jumped into bed with the Evil Empire, has he?!

"Our main goal is to make it easy to write portable application programs which work well both on DotGNU Portable.NET and on Microsoft's .NET platform."[1]

Oh, cruel fate! Nothing is sacred anymore! Now we all need to boycott GNU! Immediately uninstall coreutils, GCC, and all other GNU software! We can't support software released by a monopoly-supporting hater of freedom like the FSF! (My irony-meter has spontaneously combusted, and is smoldering in the corner).

____ [1] http://www.gnu.org/software/dotgnu/pnet.html
MonkeeSage

2008-12-21 07:55:45

Oops...original comment got lost...repost:

"...you are nonetheless betraying the principles of Free Software by supporting Microsoft’s Intellectual Monopoly..."

Then GNU / FSF are also betraying the principles of Free Software, since they have their own C#-to-CLI compiler, CLR, and CLI disassembler, just like the mono project (and you can target pnet from monodevelop, BTW).

But...that must be some kind of fluke...RMS hasn't really jumped into bed with the Evil Empire, has he?!

"Our main goal is to make it easy to write portable application programs which work well both on DotGNU Portable.NET and on Microsoft's .NET platform."[1]

Oh, cruel fate! Nothing is sacred anymore! Now we all need to boycott GNU! Immediately uninstall coreutils, GCC, and all other GNU software! We can't support software released by a monopoly-supporting hater of freedom like the FSF! (My irony-meter has spontaneously combusted, and is smoldering in the corner).

____ [1] http://www.gnu.org/software/dotgnu/pnet.html
MonkeeSage

2008-12-21 07:57:20

Oops...original comment got lost...repost:

"...you are nonetheless betraying the principles of Free Software by supporting Microsoft’s Intellectual Monopoly..."

Then GNU / FSF are also betraying the principles of Free Software, since they have their own C#-to-CLI compiler, CLR, and CLI disassembler, just like the mono project (and you can target pnet from monodevelop, BTW).

But...that must be some kind of fluke...RMS hasn't really jumped into bed with the Evil Empire, has he?!

"Our main goal is to make it easy to write portable application programs which work well both on DotGNU Portable.NET and on Microsoft's .NET platform."

Oh, cruel fate! Nothing is sacred anymore! Now we all need to boycott GNU! Immediately uninstall coreutils, GCC, and all other GNU software! We can't support software released by a monopoly-supporting hater of freedom like the FSF! (My irony-meter has spontaneously combusted, and is smoldering in the corner).
Roy Schestowitz

2008-12-21 09:36:11

He encouraged against developing in .NET. Novell promotes it..
jo Shields

2008-12-21 10:09:57

He encouraged against developing in .NET. Novell promotes it..

And if pnet only included a runtime, that might hold water.

But the existence of a compiler rather changes things, doesn't it?
Slated

2008-12-28 09:45:04

@ The spamming idiot called MonkeeSage

Ref: DotGNU:

AFAIAC DotGNU is just yet another implementation of Microsoft's .Net technology, so yes ... I do still advise people to refrain from using it, or any other encumbered Microsoft technology, wherever possible.

I support Linux (the kernel), but that does not mean I agree with everything Torvalds says (Torvalds: "I think Tivoization is *good*"). Equally I support Free Software, but that doesn't mean I agree with everything Stallman says either. IMHO Stallman's support of encumbered Microsoft technology via DotGNU, is a serious mistake.
Slated

2008-12-28 10:05:14

@Jo Shields

I believe "congratulations" are in order, since apparently you've just succeeded in poisoning Debian with Moonlight too.

Will you be joining that other FOSS traitor, Miguel de Icaza, at the next Microsoft Developers Conference, to discuss the terms of a cross-licensing "deal" with Debian, so they can purchase some Peace of Mind€®, or will you just recommend they skip straight to the litigation phase to "cut off [Debian's] air supply"?

I wonder how Debian users are going to circumvent this legal requirement for only obtaining Moonlight as direct 'Downstream Recipients' from Novell? Is Novell and Microsoft planning to merge Debian into their FOSS-assimilation portfolio?

At the rate Microsoft is assimilating Linux (with your eager assistance of course), I'd be as well giving up and buying a Mac.

Thanks for your "help".

Benchmark: Ballnux is Fat

Comments

Recent Techrights' Posts