Slashdot Mirror

If Linux is so bad, I can't seen it as a developer nor as an end user. As a matter of fact, if Linux was so bad, how come these benchmarks show Linux 2.6 outpeforming openBSD?

scalability benchmarks

Here: http://www.nvidia.com/object/unix.html
FreeBSD and Solaris drivers for nVidia cards.
About performance: http://bulk.fefe.de/scalability/

"Linux has never been about quality. There are so many parts of the system that are just these cheap little hacks, and it happens to run."

If Linux just "happens to run", how come it knocks out OpenBSD when it comes to performance? I very much doubt that Linux would win tests like these if "many parts" of its code were low quality and badly designed.

Granted, the test linked to above is soon two years old, and De Raadt refers to style of coding or general code quality rather than raw performance -- which other prominent people also have commented (in a perhaps more balanced way), but the fact that Linux runs is not merely a coincidence, as De Raadt seems to insinuate.

Does linux have fine-grained locking in its kernel?

Does linux have or support Access Control Lists?

Hmmm... where did you read that? Even in the fefe test, freebsd and linux have very similar performance characteristics, and that's a two year old benchmark.Quote:

"FreeBSD looks like it would scale O(1) if I could create more processes with it, but as long as I can't confirm it, I can only give it the second place.

"

Check the graphs ... and the corrections (author did not read man tuning, sysctl, the handbook... well, the documentation in general at first, so did not know how to set kern.maxproc).

You're welcome (to this information) :)

Please do not port software to Windows!

Uhm, what's the software firewall for?
Check out this text, I think it makes some very good points against the use of such software.

Uh, and this review is helping us... how?

lame uses assembler code for vectorization. One of the new features of gcc 4 is the beginnings of a vectorization model. A good test for gcc 4 would have been to compile some C-only bignum libraries, and Ogg Vorbis! povray is also a good example, but then you need to test more than one specific test-run. Maybe gcc 4 makes radiosity in pov-ray 400% faster at a 2% cost in the rest of the code?

This guy is the Tom's Hardware of Linux reviews, except he doesn't have the annoying ads, and he does not split his lack of content over 30 HTML pages.

The new warnings of gcc 4 have helped me find a bug in my code. That saved me a week. Consider how much faster gcc 4 needs to make pov-ray or lame to save you a week of work!

gcc 4 can now reorder functions according to profile feedback. That should make large C++ projects faster. Also, the ELF visibility should make KDE start much faster. This should have been tested!

Please note that I'm not saying gcc 4 produces faster code. I don't rightly know. I do know it produces smaller code for my project dietlibc, where size matters more than speed.

FWIW, I use Fefe's Minit, which is based on daemontools. And it was probably the first attempt at speeding up boot times in Linuxland. Good to see things like this going mainstream.

minit is exactly what you describe. With minit, I got my system's boot time from starting /sbin/init to being presented with a login prompt down to 3 seconds, with all the daemons and programs started as the normal init scripts do.

NetBSD is hampered by poor scalability and its limited rudimentary SMP.

SMP, yes at the moment. But Uni proc scalability? I don't think so. This looks interesting too.

NetBSD also lacks a production ready journaling file system.

With Soft Updates, it doesn't need it.

A bit OT, but nevertheless quite interesting to read and it contains information about SIMD instruction sets other than just MMX/SSE: http://www.fefe.de/ccccamp2003-simd.pdf

Yeah, because I *totally* pick technology based on how well I an make an *acronym* out of it.

"Hey joe!!! PostgreSQL can implement all our constraints using triggers.. we need to use this instead of MySQL, it doesn't even have views!! I can't wait to install it on our FreeBSD server. I hope it runs well with Lisp under fnord.

"Sorry Bob, those technologies spell out PFFL, or 'piffle'. I can't authorize it. Do you have any that spell out 'SOLID' or 'AWSM' or 'SECUR'"?

Linux scales better than BSD

Apache 2 and a recent Linux kernel come pretty close to the theoretical limits of the hardware when it comes to serving static content

Oh, no. You cannot have tried this in practice. Apache still suffers heavily under load. Heavy load makes response time skyrocket. Benchmark against Gatling -- now there's a fast web server!

they want to make sure everyone is nice and compliant about upgrading when they decide to take httpd over to java like all the other java kool-aid they are selling --- Maven is Jonestown, lets all program in XML because its standard! Cultures breakdown when there is too little disent and questioning of authority, the apache foundtion is headed in that direction.

Lets move on, SOA and all that, most people don't need any of this mod_* crap and could use:
thttpd he has other servers there, too and http_load.
lighttpd I'm moving to this sweet little server for most apps and the home site runs ea php and ruby on rails
AOLServer like OpenACS runs on
Boa
fnord from our boy who did the (in)famous benchmarks
Cherokee I root for this one for some reason.
gatling
cthulhu
yaws in erlang, should support more simul. connections than the unlying OS can support.
dhttpd
Litespeed check out their php benchmarks
thy
roxen
mini-httpd never tried this one
xitami I have a intranet server running for 5 yrs (without upgrading xitami) on xitami Solaris, simple, small, easy to admin, never dies max uptime was 1000 days+.
eddiefor complex load bal and geographic distribution
hiawatha

And for the love of god, please at least design your sites to get their images from images.mysite.com if possible so that you can use a non-bloatware web server to server the images, reserving horsepower on your apache server for stuff that actually _requires_ some features of apache.

http://www.hcsw.org/awhttpd/ updated on 12-06-2004
http://www.norz.org/zawhttpd.html
http://cr.yp.to/publicfile.html

they want to make sure everyone is nice and compliant about upgrading when they decide to take httpd over to java like all the other java kool-aid they are selling --- Maven is Jonestown, lets all program in XML because its standard! Cultures breakdown when there is too little disent and questioning of authority, the apache foundtion is headed in that direction.

Lets move on, SOA and all that, most people don't need any of this mod_* crap and could use:
thttpd he has other servers there, too and http_load.
lighttpd I'm moving to this sweet little server for most apps and the home site runs ea php and ruby on rails
AOLServer like OpenACS runs on
Boa
fnord from our boy who did the (in)famous benchmarks
Cherokee I root for this one for some reason.
gatling
cthulhu
yaws in erlang, should support more simul. connections than the unlying OS can support.
dhttpd
Litespeed check out their php benchmarks
thy
roxen
mini-httpd never tried this one
xitami I have a intranet server running for 5 yrs (without upgrading xitami) on xitami Solaris, simple, small, easy to admin, never dies max uptime was 1000 days+.
eddiefor complex load bal and geographic distribution
hiawatha

And for the love of god, please at least design your sites to get their images from images.mysite.com if possible so that you can use a non-bloatware web server to server the images, reserving horsepower on your apache server for stuff that actually _requires_ some features of apache.

http://www.hcsw.org/awhttpd/ updated on 12-06-2004
http://www.norz.org/zawhttpd.html
http://cr.yp.to/publicfile.html

Some open source developers disagree with you.

Open source developers usually write software because they want to use it. If they don't want to use it with Windows, then why should they port it just to satisfy some whinging Windows users? Especially the "I couldn't code my way out of a soggy paper bag, but Doing This Would Make Open Source Succeed*!" sort.

* Succeed being defined as "do what I want it to do"

A story about slow bootup and nobody mentioned minit, yet?

Well, it does exactly what some people proposed: launch stuff in parallel.
It works and is noticably faster than SysV-init (my laptop comes up in 18s).

Why do this? this defeats the purpose of free software. All it's doing is providing workarounds for people using windows and is yet another excuse for people not to dump windows. I don't see how linux would stand for this.

here is a summary that is very articulate, more so than I can put it here in a comment.

The canonical location for that benchmark is:
http://bulk.fefe.de/scalability/
not the phish-alike mirror you provided.

it appears you base these statements on this page -- http://bulk.fefe.de/scalability/ -- which appears to be 1 year old. is any fresher benchmarks and comparisons available?

Well this is just ridiculous. I get modded "troll" for dealing out harsh facts and observations, all because some mod didn't want to admit it anyway.

Just my 2 p:
Maybe you got modded Troll because somebody got tired of your bitching about an OS that introduced *a lot* of new features in this branch. The FreeBSD developers team undertook a very ambitious project, and the OS is *already* performing very nicely (here, again. It's true that any benchmark must be taken with a grain of salt, but I think it's enough to disprove whoever is oh so authoritatively stating that "FreeBSD is crap").

What's worst, you're blatantly ignoring that there's a *huge* room for performance improvement, since the Giant Lock still has to be pushed out of many subsystems. Hitting the -STABLE release means that it's production ready, not that it's as fast as it can.
Have a look at what Scott Long has to say on the development process. It could be enlightening.

FreeBSD (just as any other OS) can always use some constructive criticism. Your comments really look like puerile complaints.

--
Being able to read *other people's* source code is a nice thing, not a 'fundamental freedom'.

according to these benchmarks, linux 2.6 is faster freebsd 5.1.

Now that we have a lean userland in BusyBox and a lean C library in dietlibc, the next step is a lean kernel.

Suggestions?

Widely acknowledged fact: OpenLDAP performs extremely slow. I don't have any real benchmarks though.

Please do not port software to Windows!

It is this that makes me want to not support this guy or his game.

For one, I use Mac OSX. For two, some people don't *like* having to tweak their computer to get it to run(ala Linux). For three - some people HAVE to use Windows (for work or some lame reason like that...). For four - no one doubts the merits of Open Source until they go on tirades like this one.... It's like extremo Democrats who protest violently - they only ultimately hurt the cause.

Oops *blush*

Yeah I'm willing to give it a try. I've always liked NetBSD's philosophy better than FreeBSD's.

Besides...
"Congratulations, NetBSD! NetBSD now has better scalability than FreeBSD." http://bulk.fefe.de/scalability/#netbsd2

We'll see. I may like it better. Or not. Never know.

People wouldn't be reimplementing DJB's tools under less restrictive licensing (BSD in the case of ipsvd and runit, GPL in the case of libowfat) if he used a reasonable license in the first place, instead of giving out his software with no license to redistribute whatsoever.

I'm using runit rather than daemontools at work because we can't comply with DJB's non-license. Please be a bit more careful before calling bullshit.

It is a common mistake to think you have something faster just because you haven't yet designed in all you'll eventually need or want.

But it's also a mistake to think that all you'll eventually need or want needs to be designed in right now. It's also a mistake to put things at a more deeper layer than they need to be.

Sure the GNU libc is huge. But that's POSIX, and that's where the Hurd will bridge us from.

Funny, I'm pretty darned sure I've seen smaller implementations. Certainly, application compatibility for them isn't perfect -- but it's certainly Good Enough.

You yourself point out that the quick-and-dirty but real solution is often the one that wins in the market. Why, again, hasn't the HURD been kicked out the door with POSIX support only?

In contrast to OpenBSD, NetBSD has been the most reliable open-source OS I've used, and is the only one that runs on every single one of my systems (which are all under-5-year-old Intel and AMD PCs -- nothing exotic) without any major problems. Linux is a close second, and FreeBSD a distant third.

Based on the benchmark results here, I'm not surprised my OpenBSD experience was as bad as it was.

WARNING: This post contains what may be perceived as FreeBSD criticism and GNU/Linux advocacy. However, I have an open mind and would like to hear constructive responses to the issues I raise. I like FreeBSD, I just like Debian GNU/Linux and NetBSD better.

I wonder what FreeBSD really has to offer that makes it better than the competition.

It doesn't enjoy the popularity and mindshare that GNU/Linux enjoys. People in the server world sometimes prefer it, but outside of this world Open Source = Linux (and, usually, = Red Hat), and FreeBSD is unheard of.

It uses a much more monolithic kernel than Linux, making it lose some flexibility. You wouldn't really want to use FreeBSD for an embedded system. (And I mean really low-end, not "we call it embedded because it's small, but it's actually powerful enough to run a desktop OS".) Linux and NetBSD are more suitable choices here.

Fewer drivers are available (especially those available as binary modules for Linux). Many applications developed for the GNU system won't work on a vanilla FreeBSD system. While this is the applications' fault, it still is a disadvantage for FreeBSD. It also has fewer binary packages available than Debian GNU/Linux.

It loses against Linux and NetBSD in terms of supported architectures.

It loses against Linux and NetBSD in this benchmark.

FreeBSD systems are easy to administer using ports, but the same can be said of other BSDs. There are Linux distributions using ports (or variants thereof), and apt is at least as convenient.

So what is left? FreeBSD (and also NetBSD) definitely has a more professional feel about it than many Linux distros. It has also proven itself many times in server environments. However, with GNU/Linux (at least Debian) beating it in technical and usability aspects, are these emotions really warranted? Then there is the license. I don't think either license (BSD or GPL) can be said to be better than the other, but there must be cases where the BSD license is to be preferred, so the license could be an argument. Is that what it all comes down to, then?

PS. I've heard many people complain about the Debian installer (the one used in woody), and I've heard the FreeBSD installer time and time again. Personally, I find the Debian installer vastly superior to FreeBSD's (which has failed a number of times, and given me a hard time making the right choices, especially when partitioning). Yes, I am a Debian zealot, but let me add that the installer I've liked best is OpenBSD's.

So, what's up with the FreeBSD installer being easy to use (except for me) and the Debian installer being hard to use (except for people who read the messages it gives you)? Is it really me, or are they measured against different standards (FreeBSD being for more technically apt people than GNU/Linux)?

I assume you are talking about this: Benchmarking BSD and Linux from this slashdot story. Linux 2.6 is the clear winner in all almost all tests.

(The trick for finding it was to use google instead of slashdot search. This search found it at once.)

here is a decent benchmark I remember from a while ago. 2.6 put alot of work in and is better to on-par with most tests, though all statistics are flawed, the code for the benchmarking is open if you want to give it a shot.

sed -i -e 's/extern int errno;/#include <errno.h>/' djbdns-1.05/error.h

for i in `seq 0 127`; do ln 127.0.0.1 10.0.0.${i}; done

Isn't this the story line for some cheap porno film??

You can say a lot of good stuff about *BSD, but it currently does not match the quality and quantity of great minds work that is being put into the linux kernel.

Ivan's benchmarking document is complete.

I love to use OpenBSD, for lots of reasons. I would love top performance as a bonus, but I use OpenBSD for where I like it to fit.

I'm getting into Gentoo now, so I'm happy there too.

I don't understand why people get so upset over what performs the best. The BSD's and Linux have benefited from each other and in the end we all get to choose what fits our needs the best.

Linux 2.6 excellent performance will probably benefit the BSD's in some way, even if just urging the BSD developers on and in the future no doubt the opposite will be true.

NetBSD made great improvements due to Felix's benchmark.

Ivan's benchmarking document is complete.

I love to use OpenBSD, for lots of reasons. I would love top performance as a bonus, but I use OpenBSD for where I like it to fit.

I'm getting into Gentoo now, so I'm happy there too.

I don't understand why people get so upset over what performs the best. The BSD's and Linux have benefited from each other and in the end we all get to choose what fits our needs the best.

Linux 2.6 excellent performance will probably benefit the BSD's in some way, even if just urging the BSD developers on and in the future no doubt the opposite will be true.

NetBSD made great improvements due to Felix's benchmark.

... at packet sniffing. In other news, FreeBSD
sucks
at
everything
else

From the recent tests performed by a BSD advocate, and linked here, we see Linux 2.6 has TRIPLE the exec throughput of FreeBSD 5.2, TRIPLE the context switch speed, less than half the system call overhead, and all, of course while being far more scalable.

I suggest that if you really have the need to show that FreeBSD is better than Linux, then you should concentrate on something other than performance. Stability might suit you better, because it is far more difficult for someone to get evidence either way to counter your FUD.

Have a nice day.

The money quote from the Conclusion of Felix von Leitner's Benchmarking BSD and Linux:

OpenBSD 3.4 was a real stinker in these tests. The installation routine sucks, the disk performance sucks, the kernel was unstable, and in the network scalability department it was even outperformed by it's father, NetBSD. OpenBSD also gets points deducted for the sabotage they did to their IPv6 stack. If you are using OpenBSD, you should move away now.

http://bulk.fefe.de/scalability/

http://developers.slashdot.org/developers/03/10/19 /0130256.shtml

How many of those operating systems use Apache?

You mean including Windows, all of them can, but there are fare more webservers available for Unix like systems than there are for Windows. Thttpd, wn, Thy, Roxen, Fnord, Dhttpd, Caudium, Bozotic, Boa, and AOLserver are all available in Debian in addition to Apache. Most of these are IPv6, ssl/tls, and cgi capable. They all have their strengths, and they all are being actively maintained. Most of these will operate as a drop-in replacement for Apache for most sites.

You are correct that most of the web servers on the net are Apache installations of one type or another. Most sites do not need or use all of the features that Apache offers, but install Apache anyway. Sound familiar? They are still thinking in traditional market terms, instead of looking at what is available to them. They treating Unix as if it were Windows, but if an cross-platform Apache-specific worm were to affect them adversely, there will be alternatives available to them that they would not have on Windows.

The point is that Unix like operating systems offer greater variety of more services in more implementations than Windows does or ever will. There is more room for fault tolerance, more methods available, and more capability to find new solutions to new and old problems (including security) in Free Software than any company or group of companies is capable of providing.

I wouldn't be all that sure about your predictions.

There are many things we don't know yet:

- What is going to be measured?
- How are measurements going to be weighted to compute the final score?
- What systems are participating?

The list of participating systems is not final yet. There are other systems out there besides *BSD and GNU/Linux. I could imagine an embedded Linux (without the weigt of a full GNU userland) beating the other Linuxen.

Certain things are pretty inefficient under UNIX and like systems (e.g. monitoring file descriptors for incoming data - although the *BSDs have kqueue and Linux 2.6 has epoll, see this benchmark). A non-UNIX contender might join and beat the others.

Hardware compatibility seems to vary among Linux distros. This could also affect scores (e.g. lost time getting things to work, that could have been spent on tweaking for performance).

" OpenBSD 3.4 was a real stinker in these tests. The installation routine sucks, the disk performance sucks, the kernel was unstable, and in the network scalability department it was even outperformed by it's father, NetBSD... If you are using OpenBSD, you should move away now. "

Try this link.

It's an excellent comparison.

I'll be happy when you go troll fuckedcompany.com where you belong, you fucking coward.

Most FreeBSD users don't really care that Linux 2.6 has better overall benchmarks than FreeBSD 5.x by a cunt hair, at least according to this. It would be interesting to see these benchmarks updated now that ULE is the default scheduler...

If you were a professionial sysadmin maybe you would understand that performance is only one factor amongst many. I think I speak for most FreeBSD users when I say that Linux is an amazing OS that has obviously changed the landscape of IT for the better. We don't disagree. We simply prefer FreeBSD and would rather spend a few more bucks on fast processor and disk than put up with what we perceive as usability and maintenance headaches with Linux, its vast number of distributions and its lack of a standardized base system. However, it's not like I cringe when I have to admin, say, a RedHat system. I just have to jump through a lot more hoops to do things (like managing software compiled from source) that would be easy and standardized in FreeBSD. I'm also loath to rebuild the kernel on production RedHat machines since I'm either out on a limb without the benefit of RedHat tweaks and testing (if I grab generic source) or have to build my own kernel RPM's every week or so from RedHat's SRPM's. Those are just a couple of examples of why I'll take technically inferior but architecturally superior FreeBSD over Linux when I have the choice (and I'm talking about 4.9 here, which is obviously behind the curve, performance-wise).

Anyway, do us all a favor and FUCK OFF, you petulant brat.
*plonk*

I'll preface this by saying that I do most of my work in Linux. I prefer the Linux licensing model and the overall "experience". That said, I do lots and lots of work with FreeBSD and am active on various lists, answering mostly Unix related questions (rather that so much FreeBSD specific ones).

Here is an article about the Linux and BSD performance. I refer to this one because it matters most in what I do (deploy Linux/Unix machines for clients).

FreeBSD is good, damned good, but I think Linux is better.

OK you got 1 of 6 correct. You fail.

Linux and FreeBSD kernels, which
1. is the most SMP scalable (parallel)

Linux. (Linux and FreeBSD both started with a BKL about 5 years ago. Linux is now being used on 512 processor machines, FreeBSD doesn't scale past 4 CPUs for all their research).

2. is the most algorithmically scalable
Linux

3. has fastest single threaded performance
Linux (from the horse's proverbial mouth).

4. runs on more architectures
Linux. See here and here (The Linux Kernel supports more architectures than the NetBSD kernel, idiot).

5. supports the most hardware
Linux. See here and here (NetBSD is actually the one that prides themselves as running on toasters. And nobody cares about your shitty DEC Alphas. IA64, POWER are where its at now).

6. has the fastest TCP/IP stack
Linux

See you again next year.

Maybe in your experience. But, my experience with installations of HP/UX, AIX, SunOS/Solaris in the largest entertainment companies based in SoCal and at least one of the largest Japanese electronic companies with offices in SoCal use GNU utils. In addition, others appear to disagree with you: "Most Unix systems today include many tools from the GNU project simply by default, even though they are not necessarily GNU systems!"

At least the GNU utilities made a uniformly bad implementation available across all platforms...

Bad enough to be used in the Open Source BSD distributions.

First, they did anything but lock it away. As I pointed out, Sun opened up Unix

Really?: " Proprietary versions of Unix were becoming popular in the corporate arena, and quite often these versions lacked source code (making them nearly unusable). Sun even distributed versions without C compilers! The personal computer was taking off, championed by proprietary software vendors Microsoft and IBM. If not for GNU, some argue, this disturbing trend of proprietary operating systems might have become the standard."

And NFS was an open spec from day one, and always available for licensing even to competitors on commercailly attractive terms

Affordable by developers like myself? Ha! Really, the wonder of it all. It's easy to be flippant about the "attractive terms" a license would be on a corporate budget, but we're talking about how individuals changed the Unix world by creating their own Open Source tools rather than something an individual couldn't afford. And, even if an individual Open Source developer could afford it, he'd be in the minority amongst his peers.

Sun opened up Unix, and was single handedly reponsible for proving that it was a viable alternative to the IBM and VAX machines that ruled in those days.

Sun was part of the Unix Wars, and didn't single-handedly do anything but make its own version of BSD proprietary. Yes, they did add functionality into their proprietary version which made it a top contender in the Unix Wars. As for NFS, Sun just doesn't know what to do in order to ingratiate themselves with Open Source developers and still maintain strict control over how people use their "Open" spec. But, clearly the market is dragging their asses out of the closet.

Sure, there are lots of Linux fan-boys who claim this, but it's simply not true.

At least one Unix vendor disagrees with you: "as most Linux kernel testing efforts have only been conducted over short periods of time, this series of tests provides us first-hand data and results of longer runs. The series of tests also provides data for heavy-stress workloads on Linux kernel components, as well as TCP, NFS, and other test components. The tests demonstrate that the Linux system is reliable and stable over long durations and can provide a robust, enterprise-level environment." And, other researchers disagree with you as well:"FreeBSD has by far the best performance of the BSDs and it comes close to Linux 2.6". Naturally, you can continue to believe that it's not true by disregarding the facts, but then you're simply operating on faith, not reason.

BSD's advantages

We're talking about Linux vs FreeBSD here, clown, not to mention that you provided zero evidence. You got failed.

Linux and FreeBSD kernels, which
1. is the most SMP scalable (parallel)
Linux

2. is the most algorithmically scalable
Linux

3. has fastest single threaded performance
Linux

4. runs on more architectures
Linux. See here and here

5. supports the most hardware
Linux. See here and here

6. has the fastest TCP/IP stack
Linux

Hope this has been helpful. See you again next year.