Measuring The Benefits Of The Gentoo Approach

An anonymous reader writes "We're constantly hearing how the source based nature of the Gentoo distro makes better use of your hardware, but no-one seems to have really tested it. What kind of gains are involved over distros which use binary packaging? The article is here."

Comments

Misses the point

[keesh]

The source-based thing isn't even why most people use gentoo. According to a recent poll on the gentoo-user mailing list, most people like it because of Portage (the package management system), with Customisation / Control coming in second (performance was third). Portage rocks. Even with the compiling, it takes less time to install some stuff (eg nmap) than it would take to locate the relevant .rpm. Of course, kde's a different matter, but with distcc compiling doesn't take too long.

Having said that, it looks like the guys doing the testing got their CFLAGS wrong. Gentoo's performance should never be worse than Mandrake -- I reckon they forgot omit-frame-pointer. Also, the kernel compile is unfair, because gentoo-sources includes a whole load of patches that Mandrake and Debian don't.

Finally, what's with measuring compile times? How is that a fair way of measuring performance? Hey, look, my distcc + ccache + lots of CPUs system with gcc3.2 can compile stuff faster than your single CPU gcc2 system... It's like comparing chalk and oranges.

Re:Misses the point

[ecchi_0]

Finally, what's with measuring compile times? How is that a fair way of measuring performance? Hey, look, my distcc + ccache + lots of CPUs system with gcc3.2 can compile stuff faster than your single CPU gcc2 system... It's like comparing chalk and oranges

Except in this case they all had the same hardware on each machine...

Re:Misses the point

[Anonymous Coward]

ccache and distcc are software, and they can make a huge (as in several orders of magnitude for things like mozilla) difference

Re:Misses the point

[arkanes]

The key points to recognize from the article are: a) GNUmeric's performance sucks (8 minutes to open a file? I won't even think about the other version...) and b) that the CPU is not a signifigant bottleneck in modern systems. We all knew that. It's one reason why so many people are happy with binary packages, because the speed increase from saving some cycles generally isn't worth the extra time you lose compiling (as seen, in many cases it makes 0 difference).

I would have liked to see some tests with things that are more CPU than IO bound, but, realistically, how often do you do those things in the normal case?

If the main reason is to use portage for the convenience (same reason many people use debian), maybe they need to expand portage to support binary packages.

Re:Misses the point

[tweek]

Well in all fairness, ccache only does any good after the first compile. The distcc option however does make a difference.

I will agree that the biggest thing for me with gentoo is actually being able to strip stuff out of an install with a simple USE flag. I actually prefere to build things myself but having a package management system that takes care of dependencies for that is a godsend.

Re:Misses the point

[aboyce]

first, I'd love to see a distro be faster than "up2date package_name" or even "aptget package_name".

Next, they said right in the article that they used an identical copy of the kernel source on each machine, so patches shouldn't make a difference.

Finally, its not that I dont agree with you, their tests did have flaws, it just seems that some of your facts are wrong in attacking them. There are some points that need to be examined, even if some of their conclusions are premature.

Re:Misses the point

[scotch]

Why the hell would you introduce a distributed computing tool in a discussion about evaluating the performance of a single machine/OS? Pure obfuscation. Typical gentoo-missing-the-point behavior. Either compiling the kernel is a fair measure of the speed of the system or it isn't. distcc doesn't play into it. Here's a fun analogy. You want to see which is faster a porche 911 or a chevy corvette. So some thoughtful guys put together a series of tests, one of which is a 1000 mile race. Then along comes user keesh (202812) who says "Bad test, I wouldn't drive 1000 miles, I would take the train."

A hearty helping of wtf is in order. Some of your other points are ok, though ;).

Re:Misses the point

[Anonymous Coward]

emerge is faster than up2date. one character less to type.

Re:Misses the point

[countvlad]

Portage can be used to install binary (precompiled tbz2 packages of ebuilds).

From emerge --help:

--usepkg (-k short option)
Tell emerge to use binary packages (from $PKGDIR) if they are available, thus possibly avoiding some time-consuming compiles.This option is useful for CD installs; you can export PKGDIR=/mnt/cdrom/packages and then use this option to have emerge "pull" binary packages from the CD in order to satisfy dependencies.

--usepkgonly (-K short option)
Like --usepkg above, except this only allows the use of binary packages, and it will abort the emerge if the package is not available at the time of dependency calculation.

You can also, of course, emerge rpm and install any RPM packages. I'm not sure about debian .deb packages or slackware .tgz packages.

Gentoo is also accept pre-orders for it's upcoming 1.4 release. Information can be found here, at the Gentoo Store.
They even have precompiled packages optimizaed for Athlon-XP's - drool!

Re:Misses the point

[raventh1]

They used the same kernel source. says so in the article...

Re:Misses the point

[mickwd]

"emerge is faster than up2date. one character less to type."

Hehehe....and Mandrake's "urpmi" is one character shorter than that.....

Re:Misses the point

[Anonymous Coward]

Not a good troll. Good trolls are subtly misleading and well written. Yours isn't. By far.

Re:Misses the point

[FeeDBaCK]

Portage has supported binary packages for a while now and the current beta version of portage even has support for automatically downloading binary packages.

Re:Misses the point

[dmon]

> Even with the compiling, it takes less time to install some stuff (eg nmap) than it would take to locate the relevant .rpm

Gentoo is really cool, but:

http://apt.freshrpms.net/

# apt-get install nmap

that's it.

Re:Misses the point

[keesh]

alias e sudo emerge. I win.

Re:Misses the point

[antiMStroll]

What's missing in the article is the second half of Gentoo's compile options, the /etc/make.conf USE variables. CFLAGS determines CPU architecture, USE adds or removes the options for extra software support. In stock form Gentoo compiles binaries with a huge number off add-ons, including support for KDE, Gnome, framebuffer, etc. From make.conf: " USE options are inherited from /etc/make.profile/make.defaults." The list from a current Gentoo 1.2 looks like:

USE="x86 oss 3dnow apm arts avi berkdb crypt cups encode gdbm gif gpm gtk imlib java jpeg kde libg++ libwww mikmod mmx motif mpeg ncurses nls oggvorbis opengl pam pdflib png python qt quicktime readline sdl slang spell ssl svga tcpd truetype X xml2 xmms xv"

Without knowing what support Debian or Mandrake used to compile binaries, this is still an apple/oranges comparison. My notebook isn't configured to compile with KDE or Gnome extensions because the hardare is too old and I use Fluxbox. Mandrake and Debian may still turn out faster (the Gentoo Mozilla e-build was legendary for being slow), but that's not quiet yet proven.

Re:Misses the point

[FeeDBaCK]

Either compiling the kernel is a fair measure of the speed of the system or it isn't.

It definitely is not unless they were using unpatched sources in all three systems. The Gentoo sources applies bunches of patches to the stock kernel which would affect compile time.

Also, am I the only one that noticed they complain about how all the speed improvement patches made no difference, but they mention that they used the same options for every kernel compile. The performance increases are found by TURNING THEM ON.

To go with the car analagy, it would be like having a 6th gear in the Porsche but only using 5th gear because the other one didn't have it.

I udnerstand using the same options for TIMING a kernel compile, but not for running one. After all, the ability to customize your system is what makes Gentoo so nice.

Re:Misses the point

[Anonymous Coward]

Uhm... with emerge, it's just 'emerge nmap'. That handles dependencies as well if necessary.

Re:Misses the point

[scotch]

'chsh -s /bin/emerge' I win.

Re:Misses the point

[rf0]

Of course if the do a build, kill it half way through then build it again with out make clean it will go even quicker. Not saying they did but stats can always be taken two ways

Rus

Re:Misses the point

[cperciva]

first, I'd love to see a distro be faster than "up2date package_name" or even "aptget package_name".

FreeBSD Update. Ok, it only upgrades the base FreeBSD install, starting at binary releases, along the security branches; but it uses binary patches to dramatically cut down on the bandwidth usage (and therefore the time used). A typical install of FreeBSD 4.7-RELEASE (released in October 2002) has 97 files totalling 36MB bytes which need to be updated for security reasons; FreeBSD Update does this while using under 1.6MB of bandwidth.

Re:Misses the point

[buchanmilne]

Portage rocks.

If you have a fast processor. My Duron 800 can keep itself busy for a weekend compiling OpenOffice.org ...

Even with the compiling, it takes less time to install some stuff (eg nmap) than it would take to locate the relevant .rpm.

This is on my Thinkpad 600X, which is a 500 PIII/192MB, with a pretty slow disk:

[root@bgmilne-thinkpad mnt]# rpm -q nmap
package nmap is not installed
[root@bgmilne-thinkpad mnt]# time urpmi nmap
installing /var/cache/urpmi/rpms/nmap-3.00-2mdk.i586.rpm

Preparing...
#some hashes replaced to fool the lameness filter#
1:nmap
#some hashes replaced to fool the lameness filter#
5.34user 1.36system 0:26.76elapsed 25%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (1712major+8394minor)pagefaults 0swaps


You would need quite a system to beat 26s I think.

Also, the kernel compile is unfair, because gentoo-sources includes a whole load of patches that Mandrake and Debian don't.

From the article:

"The same 2.4.21 source was copied to all machines and compiled using the same options. However, it should be noted that the Debian system used gcc 3.3.1 whilst the Mandrake and Gentoo installations used gcc 3.3.2 ."

I don't see the point of:
-not using the default compiler on the system
-if you don't use the default compiler on each machine, at least use the same compiler across them all

But, otherwise, the comparison looks pretty fair.

Re:Misses the point

[Vellmont]

I think you're probbably right about the reasons for using Gentoo (though I admit I've never actually used Gentoo).

However, I think that a kernel compile _is_ a fair measure of overall system performance. It involves lots of disk, memory, and processor access, so it's a decent indicator of across the board performance.

As far as kernel compile versions go, from the article:

The same 2.4.21 source was copied to all machines and compiled using the same options. However, it should be noted that the Debian system used gcc 3.3.1 whilst the Mandrake and Gentoo installations used gcc 3.3.2 .

So the kernel source was the same, not Gentoo source.

You say the performance problems are because they got the CFLAGS wrong. If this is the case it only seems to underscore how easy it is to screw up optimizations with Gentoo. It's great for people that know all the proper optimizations for a particular piece of hardware, but I think the majority of people just don't know this offhand.

In any case I find it very interesting the big differences you can see in performance between distributions on the same hardware (and I'm assuming similar kernel versions).

Re:Misses the point

[lspd]

It definitely is not unless they were using unpatched sources in all three systems. The Gentoo sources applies bunches of patches to the stock kernel which would affect compile time.

RTFA. "The same 2.4.21 source was copied to all machines and compiled using the same options. However, it should be noted that the Debian system used gcc 3.3.1 whilst the Mandrake and Gentoo installations used gcc 3.3.2"

Re:Misses the point

[Sancho]

One of my favorite uses for Gentoo is optimizing for size rather than execution speed. As you say, the CPU is rarely the bottleneck these days, but loading files from the disk can be a factor in a program starting up. I've done the benchmarks, and some rather large programs see significantly reduced load times when optimized with -Os.

Re:Misses the point

[dougmc]

the CPU is not a signifigant bottleneck in modern systems.

Are you on crack? Even today, the CPU speed is a signifigant bottleneck for many operations.

Now, the cpu speed has increased by a larger factor than memory speed and disk speed over the last few years, but it's still quite a large bottleneck in many operations, including the ones tested in this (admitedly lacking) test.

Even the speed of a kernel compile, which is often given as a classic `disk I/O bound' process, is extremely CPU bound. How do I know? Running `top' on an idle box shows 0% cpu utilization. Once I start the compile, it goes to well over 90% cpu utilization and stays there until the compilation is done. (just to be complete, I'm testing this on a dual p3 700 box with SCSI disks, doing a `make -j2'. But even my 2ghz Athlon computer with IDE disks works similarly.)

Perhaps I'll do some tests with adjusting the CPU multiplier on a given box, see how that affects compilation times. That would be an excellent test ...

Re:Misses the point

[Anonymous Coward]

"...they said right in the article that they used an identical copy of the kernel source on each machine..."
Where do they say that? It looks to me like the gentoo kernel was different. Near the end of the article: "Likewise, the stock gentoo-sources kernel includes optimisations for interactive desktop usage but in our (limited) user impressions this benefit did not show through." I havn't used gentoo, so I couldn't say if that actuall indicates patches or just compile options, but I don't get the feeling it was the same kernel as the other setups.

Re:Misses the point

[Anonymous Coward]

Portage rocks. Even with the compiling, it takes less time to install some stuff (eg nmap) than it would take to locate the relevant .rpm.

Honestly, that's just because Red Hat sucks. Debian doesn't have that problem. apt-get install nmap. Done. I recently had to build a Red Hat 9 system and it felt like I was traveling back into the stone age. Even Mandrake has urpmi for god's sake. Plus updates! I have to sign up and pay for updates from Red Hat if you want to use up2date. Otherwise you have to manually do it apparently. Debian + Mandrake don't have that problem again. Red Hat == old and busted. Debian/Gentoo/Mandrake == new hotness.

Re:Misses the point

[Anonymous Coward]

Oh, I see, the kernels used for the compile time test were exactly the same -- my bad.

Re:Misses the point

[nagora]

but, realistically, how often do you do those things in the normal case?

You obviously don't use GIMP or analyse OS mapping data much. I have the RAM to get the info into memory, IO is not an issue for much of my work.

Having said that, portage is the main reason I've converted all my machines to Gentoo; it's just not a serious option to go back to RPM based systems after using it for a week or so.

TWW

Re:Misses the point

[gibber]

Changing the CPU multiplier will not give you what you are looking for as it will likely change your FSB speed and L1/L2 cache access rates. The most common bottlenecks on systems are heirarchical IO bandwidth related. For example, having scads of RAM for buffer/cache will help you with disk IO woes.

Compiling binaries with optimization for a particular processor help with i-cache and d-cache utilization. The fewer instructions fetched (or the order in which they are fetched) makes a big difference in performance.

Boosting CPU cache size (up to practical cache limits), increasing FSB speed and avoiding disk IO are much more significant than CPU M/GHz.

Compilation, especially optimized (-0X) compilation is _VERY_ CPU intense. If you have enough RAM to avoid the disk thrashing caused by writing numerous intermediate files you will peg your CPU.

Most user activities (aside from games) on computers are not bottlenecked by CPU but by various heirarchical IO constraints and hence the previous poster was correct that the CPU is not a significant bottleneck on modern systems.

Re:Misses the point

[tau_]

"Having said that, it looks like the guys doing the testing got their CFLAGS wrong. Gentoo's performance should never be worse than Mandrake -- I reckon they forgot omit-frame-pointer."

That's a big reason why I'd not switch from a binary-distributed OS (I happy with Red Hat, YMMV) to a source-distributed one. Not only is it more convenient to get ready-to-run binaries, but I can actually trust the vendor to have tried to choose a good set of compiler options and to have tested the results, so I don't have to.

Any programmer worth being called one could tell you that -O2 -march=i686 isn't even the beginning of looking for the best optimisation for any piece of software. Dozens of flags affect the output, sometimes there are choices to be made between size vs speed, sometimes "optimize for size" actually produces the fastest code, and so on and so on. And some combinations actually produce buggy results. It's a huge amount of work to find the optimum setting, and the CPU model specific scheduling tweaks enabled by -march=i686 rarely make any difference at all.

I'll stick with binaries someone has tested to work, thanks. Only when I have linking problems or a feature I want had not been included at compile-time do I touch source code.

Re:Misses the point

[arkanes]

You're right, I don't, and neither does anyone else. Even in your case, memory IO is likely to be as much of a bottleneck as the CPU, if not more. One of the reasons the new Mac walks over PCs in Photoshop benchmarks is massive memory bandwidth.

Re:Misses the point

[Squinky86]

The compiling was done the same for the kernel, but the packages weren't. The precompiled packaged binaries of the other 2 are not the same as the compiled package on the gentoo system. That having been said, compiling with arch=p4 and O3 will give you much better results than the biased test would have you believe.

Re:Misses the point

[unixbob]

The source-based thing isn't even why most people use gentoo. According to a recent poll on the gentoo-user mailing list, most people like it because of Portage (the package management system), with Customisation / Control coming in second (performance was third).

Having recently tried gentoo I have to say that it's a nice idea but it still needs work. For a month I tried emerge kde. Every single time it compliained about XFree86 not matching my md5sum. So every few days I would emerge sync then emerge kde to see if the portage for XFree86 matched the md5sum I had downloaded. I removed the contents of /usr/portage and resynced but to no avail. I posted help messages on the forums of the gentoo.org site, but no one seemed to either

a) know the answer to my problem
b) recognise that there was a problem to be fixed.

I was perfectly able to install other software such as postgres, apache2, php. The idea is a good one, and watching your sources download and compile to your own specification is definetly neat. But to leave such a glaring bug for so long just frustrated me and eventually I tired of waiting and went back to my regular distro. I can understand that some may say that I should stop whinging, get off my ass and fix the problem myself. Well maybe so. And if I had been using gentoo for longer and was more familiar with it and the people who run it, then I may have done. But as an intrigued newcomer all I can say is, close but no cigar.

Re:Misses the point

[be-fan]

I second that. I don't use Gentoo because I think I can get a minscule 0.5% extra performance by compiling myself, but because of Portage, and the Gentoo community. Portage is awesome, and the source-based nature means that ebuilds come out extremely quickly, and are less subject to distro-specific "customizations" (read: quirks) than binary packages. All the ebuilds BreakMyGentoo.net, as well as the ebuilds posted to the bug tracker are a phenomenal example of how the power of portage allows a relatively small community to maintain a large software library. The Gentoo community is also one of the nicest out there (forums.gentoo.org, rocks).

Re:Misses the point

[Dasher42]

But if you look at a large project, you'll often header files and other #include'd objects compiled over and over. ccache is a godsend, I wish I'd had it years ago.

Re:Misses the point

[loginx]

I've read the article very carefully and I've also looked at the people who wrote it...

Real-world benchmarks is a serious matter, I don't even know why this article made its way into /.

It's rather obvious that those people were not familiar with CPU optimizations and were not thorough with the benchmarking considering they didn't even bother to check for the version/revision of the base package of the distros they were working with even though they do admit that even minor revisions play a considerable role with performance.

1) What are the version/revisions of GCC on those machines? A package compiled and optimized with GCC 3.3 or even the 3.4 beta will obviously offer much better performance than an old deprecated gcc 3.2 that will be installed by gentoo by default if you haven't set up your ACCEPT_KEYWORDS

2) What hard-drive optimization did they set-up? Distros like Mandrake, Debian or Redhat set up HDParm optimizations after the first install, gentoo barely does... That would already make a big difference while opening a 32,000 lines spreadsheet...

3) What are the versions/minor revisions of the Gnome window manager on all those boxes? and GTK? Those packages provide the controls and rendering for Gnumeric... having any difference in these is not fair-play either... (try to install the same version of Gnumeris on a redhat 9.0 and a redhat 7.2 and see if the performance is the same)

To get back on the example of the sports-car race, this is kind of like benchmarking a porche and a ferrari, but you put diesel in the ferrari and forget to inflate your tires...

Basically, if you don't have experience using gentoo on a system for a while and know how to optimize your system, don't go and say that the optimizations don't work... they work perfectly well for me but I couldn't see a difference in my first 2 weeks of using gentoo... it's something you have to learn... you can't just install it and pretend the distro will self-optimize for you...it's not even supposed to.

Re:Misses the point

[realdpk]

Try -static - on FreeBSD it can improve performance for very fast running binaries significantly (such as ones designed to run 100s of times/s). I dunno about on Linux (although I've noticed that Linux seems to prefer dynamic binaries for everything.)

Re:Misses the point

[FeeDBaCK]

Replying to myself to point out that what countvlad said above is true, but there are also the undocumented -g and -G funtions of emerge which are only available in the latest beta versions of portage for inclusion in the next stable release. These functions allow for automatically downloading the .tbz2's, rather than them having to already be in ${PORTDIR}/packages before starting emerge.

Re:Misses the point

[realdpk]

'linux init=/bin/emerge' I win. ;)

Re:Misses the point

[Anonymous Coward]

Yes, [debian] and [debian] we [debian] all [debian] know [debian] that [debian] rpm [debian] and [debian] gentoo [debian] are [debian] the [debian] only [debian] two [debian] choices [debian].

If [debian] only [debian] there [debian] were [debian] a [debian] GNU/Linux [debian] distro [debian] that [debian] had [debian] excellent [debian] quality [debian] control, [debian], ran [debian] on [debian] a [debian] huge [debian] number [debian] of [debian] hardware [debian] platforms, [debian], and [debian] had [debian] over [debian] 9,000 [debian] packages [debian] available [debian] for [debian] it [debian].

Re:Misses the point

[Anonymous Coward]

u missed the point. todays systems bottelnecks are the FSB or memory clocking and the command/read/write latency. hypertransport needs to be implemented in more processors.... and other embedded chip designs.... (other than AMD)

Re:Misses the point

[bhtooefr]

Correct, but stop with the flamebait, will you? The CPU multiplier is a number that is multiplied by the FSB speed to get the CPU speed.

Re:Misses the point

[Deadplant]

CPUs still aren't fast enough for me.
maybe I'm not representative of 'most users'

at work I do video processing and at home i play games and encode DVDs to mpeg4.
Even 2.4ghz cpus take hours to encode entire movies... I can get a little better than realtime encoding to mpeg4, but when you add two-pass, and the fact that the videos are so damn long... I wish i could get a terahertz CPU...

you try running a 2 pass encode on a 6 hour 720x480 DV video file and then tell me your CPU is fast enough.

it always makes me chuckle when people say that 'this or that' part of the PC is as fast as it needs to be.
There all sorts of things we can't do because our systems aren't nearly fast enough.

Re:Misses the point

[Arker]

Actually I think the key sentence of the article was this:

The Gentoo setup by Bill Kenworthy was compiled using the "stock" kernel source and the "-march=pentium3 -pipe -O3" compile flags.

Doh! No wonder it sucked.

O3 turns on things like inlining that are only worthwhile in certain circumstances, but are often counterproductive. So the results aren't surprising in the least.

Re:Misses the point

[i am fishhead]

In many cases, gentoo packages will remove compiler options that have been found to produce buggy code, so this is usualy not a problem.

Re:Misses the point

[Anonymous Coward]

Gentoo running 3.3.2? I'm pretty sure you're wrong about that:

# emerge sync
[snip]
# ls -l /usr/portage/sys-devel/gcc
total 232
-rw-r--r-- 1 root root 25428 Jul 24 12:16 ChangeLog
-rw-r--r-- 1 root root 7246 Jul 24 12:16 Manifest
drwxr-xr-x 9 root root 4096 Aug 2 18:17 files
-rw-r--r-- 1 root root 5464 Feb 13 09:29 gcc-2.95.3-r7.ebuild
-rw-r--r-- 1 root root 9250 Jun 25 13:21 gcc-2.95.3-r8.ebuild
-rw-r--r-- 1 root root 8766 Feb 13 09:29 gcc-3.0.4-r6.ebuild
-rw-r--r-- 1 root root 9814 Jun 12 16:12 gcc-3.1-r8.ebuild
-rw-r--r-- 1 root root 9648 Jun 12 16:12 gcc-3.1.1-r1.ebuild
-rw-r--r-- 1 root root 12936 May 16 18:18 gcc-3.2-r5.ebuild
-rw-r--r-- 1 root root 14840 May 16 18:18 gcc-3.2.1-r7.ebuild
-rw-r--r-- 1 root root 15119 Jan 11 1904 gcc-3.2.2-r2.ebuild
-rw-r--r-- 1 root root 15327 May 27 17:41 gcc-3.2.2.ebuild
-rw-r--r-- 1 root root 16423 Jul 20 12:42 gcc-3.2.3-r1.ebuild
-rw-r--r-- 1 root root 16418 Jul 23 17:13 gcc-3.2.3-r2.ebuild
-rw-r--r-- 1 root root 15543 Jul 20 12:42 gcc-3.3-r1.ebuild
-rw-r--r-- 1 root root 15744 Jul 20 12:42 gcc-3.3.ebuild
#

Re:Misses the point

[Anonymous Coward]

if only I had the mod points to mod this up.

Re:Misses the point

[Anonymous Coward]

dd if=/bin/emerge of=/dev/hda1

I.. uh.. nevermind.

Re:Misses the point

[gilesjuk]

What's slower, main memory and disk access or cache memory?

Cache is fast, if you shrink code so much that it uses cache better then you speed up many operations.

Of course it depends on what you are doing, databases need a different approach to 3D rendering.

Re:Misses the point

[Anonymous Coward]

if only I had the mod points to mod this down.

Re:Misses the point

[Evilive]

The source-based thing isn't even why most people use gentoo. According to a recent poll on the gentoo-user mailing list, most people like it because of Portage (the package management system), with Customisation / Control coming in second (performance was third). Portage rocks.

Exactly why I use it.
Being able to keep out unwanted fluff libraries (alsa, mmx, other things my pc or I have no need for) or easily include things like GTK2 when the program supports that compile time option, etc.
It took a while to compile big stuff (XFree, for example)on my PII 333, but before, when I had a 1Ghz with 512M of RAM shit really rocked.
Even with my slower machine, I still started with a stage 1 tar ball.
It was worth the wait.

Re:Misses the point

[Xabraxas]

the gentoo-sources kernel is not a stock kernel. It is highly patched. Vanilla-sources is the stock kernel. I've compiled both and vanilla is much quicker.

Re:Misses the point

[Anonymous Coward]

You may have downloaded from a mirror that is slow to update. The downloaded XFree file can be deleted from /usr/portage/distfiles/

Re:Misses the point

[surprise_audit]

I didn't see anything in the article about unpacking the kernel source. Gentoo does it every time, and patches it too, before compiling it. If the other distros started from an unpacked kernel source tree, there's a difference right there...

Re:Misses the point

[p00ya]

> reckon they forgot omit-frame-pointer

It probably wouldn't have hurt them to add -fforce-addr -fforce-mem while they were there. I'm not sure about -O3, I've had varying results with it from system to system and version of gcc. The best way to check which is more appropriate out of -O3 and -O2 is look at the generated assembly, see what differences there are and get a feeling for how well its going to run (think branch prediction and cache misses).

As for the test, I think they could have afforded to provide a little more detail into the setup and how they were actually measuring the time.

Either way, I'm happily running debian with a -ck kernel; it doesn't have all the gentoo niceness in it by default, but apt is about as nice as portage and doesnt have the drawback of having to wait 10 minutes for a package of passing interest to compile.

Re:Misses the point

[Arker]

Seriously though, doubling the access speed of your RAM is likely to do more good on that sort of task than doubling the CPU speed.

Re:Misses the point

[dh003i]

Yes, but GCC 3.3.2 compiles things slower than GCC 3.3.1. Now, you can say, "that's gentoo's fault for using GCC 3.3.2". Wrong. It's the user's fault. Gentoo allows you to choose what version you want.

Re:Misses the point

[afroborg]

If only I had a gun...

Re:Misses the point

[Anonymous Coward]

Yeah, yeah. Whatever. I stopped listening to Gentoo fanboys a long time ago.

Re:Misses the point

[RzUpAnmsCwrds]

"One of the reasons the new Mac walks over PCs in Photoshop benchmarks is massive memory bandwidth."

Actually,
No.

Canterwood systems have a Pentium 4 with the 800 megahertz FSB and Dual DDR 400. Dual DDR 400 happens to have a total theoretical bandwidth of 6.4 Gigabytes a second, exactly the same as the P4's front-side bus. Plus, the busses are running synchronously so latency is lower.

The Apple G5 has a 1Ghz FSB and Dual DDR 400. Hmmm... that extra 200 mhz of FSB doesn't really do much, does it? It's still limited by memory bandwidth. Moreover, that bandwidth is shared between both processors in a DP system.

So, no, the G5 doesn't walk all over PCs in Photoshop because of two reasons:

- Photoshop is more optimized for the Mac. Adobe designed the filter algorithms to run fast with Altivec and the PPC architecture, therefore, they are inefficent when run on an x86 processor
- Apple ran the test, and therefore they chose which filters to run with which settings and how many times on which images. Clearly, Apple would choose tests that benefit them. According to their results, they used a 650MB test image.

Here's why Apple's numbers don't mean shit:

- They don't give the system configuration of the PCs. Or the Macs. We have no idea if they loaded the Mac with 8GB of memory and left the PC with 256MB. Hell, the default configruation of the Dell system with the 3Ghz P4 is 512MB of DDR - the high-end PowerMac (which they undoubtibly used) has 1GB. So, if they used the systems stock (as I suspect they did), the 640MB Photoshop image they did the test on would fit in memory on the Mac but have to be swapped on the PC.
- Assuming 32-bit RGBA color on the image, that's a 160 MEGAPIXEL image. Why would they use an image so large? Perhaps it's because it fits neatly inside the PowerMac's high-end configuration's default memory allocation but not the high-end Dell's default configuration. Makes you wonder.

The P4 with 800mhz FSB has JUST AS MUCH MEMORY BANDWIDTH as the fastest dual G5 system.

A 2-way Opteron system, on the other hand, has Dual DDR 400 for each CPU. Twice as much bandwidth as the fastest Apple system. Of course, you need a NUMA aware OS to take advantage of this bandwidth.

Re:Misses the point

[Anonymous Coward]

-Os with shared objects is more about reducing both page faults and cpu cache misses by reducing size for situations where many applications are running on cheap hardware and as you said -static is rather usefull for one application eating all cpu

Re:Misses the point

[GlassHeart]

it looks like the guys doing the testing got their CFLAGS wrong. Gentoo's performance should never be worse than Mandrake

Makes you wonder how many Gentoo users actually get their compiler flags right, doesn't it?

Re:Misses the point

[Anonymous Coward]

U r dum

k'thnx

Re:Misses the point

[groomed]

You are basically arguing that Gentoo offers similar performance to Debian, provided Gentoo is tweaked to the point where it is almost identical to Debian. But at that point, why not just install Debian??

Re:Misses the point

[Natalie's Hot Grits]

Seriously though, its not. MPEG4 encoding is SERIOUSLY cpu dependant, not bandwidth. Just look at the benchmarks of the opteron vs Pentium4 on 333MHz memory bus and you will see what I mean (in mpeg4 encoding). i'm not saying the opteron is inferior for mpeg4, but the fact that MPEG4 (at least on windows) processing is highly SSE2 optimized for the P4.

(hint: P4 beats the opteron, even though the opteron has a higher real world memory bandwidth and less memory latency and larger cache)

Yea, memory bandwidth is the bottleneck in just about everything every day users are doing. But for specific tasks such as video codecs, CPU speed is a huge bottleneck. Moreso than memory bandwidth.

Re:Misses the point

[xophos]

And if you are looking for a real performance difference try python under gentoo with full optimizations in comparison to the standard debian package. The gentoo python is up to 5 times faster, depending on the python program you feed it.

Re:Misses the point

[Anonymous Coward]

Well, emerge is my BIOS - so I win!

Re:Misses the point

[BJH]

RH doesn't require you to pay for updates - you get one 'free' machine per registered RHN user. So just do what I do - register a new user for each machine ;)

Re:Misses the point

[Anonymous Coward]

You would need quite a system to beat 26s I think.

Yeah installing rpm's once you have them takes no time at all. It's getting them that sucks. I don't know how many rpms I download and find out that they have been compiled against a different version of gtk or the kernal modules have different CRC's or even a different glibc (God, I thought we were done with that problem 4 years ago). And the vast majority of the time the packages doesn't need the new version, it just so happened that the developer was playing around with some CVS version of the library when he packaged the software.

It is not at all uncommon for me to spend an hour trying to track down all the dependencies for an rpm, and end up just downloading the source, which compiles and installs just fine.

I have tried several linux distro's (redhat 4, slackware 4, redhat 6, debian (potatoe?), mandrake 9.1) and I have to say that slackware was actually had the least amount of headache in installing packages. Compiling everything did take the computer longer, but it saved me time compared to rpm's.

Debian may be nice but it took me a week to get installed since I had to install a 2 year old version that did not recognise much of my hardware and then apt-get to upgrade the entire system to a new version. Then a week later my deskstar harddrive failed, and I didn't feel like going through that install process again.

I think gentoo will be my next distro to try. For the longest time I dismissed it because everyone I knew that ran it liked it because it was "custom optimised" for their hardware, which I knew was crap. So I tended to look at them like I do audiophiles. It wasn't till recently that I realized gentoo's potential for easing the rpm-dependency hell that other distro's have.

Re:Misses the point

[More Karma Than God]

Can't you install Portage on any distribution?

If so, why not use it on Mandrake or one of the other more newbie-friendly distributions.

Re:Misses the point

[cha0sadddddddd]

Debian may be nice but it took me a week to get installed since I had to install a 2 year old version that did not recognise much of my hardware and then apt-get to upgrade the entire system to a new version. Then a week later my deskstar harddrive failed, and I didn't feel like going through that install process again.

So just for your future reference...

download knoppix iso
at boot prompt type knoppix 2
when done booting type knx-hdinstall

so please dont talk about how you HAD to install a 2 year old version.

A semi-retarded monkey could install debian using a knoppix cd.

Re:Misses the point

[Halo1]

hotoshop is more optimized for the Mac. Adobe designed the filter algorithms to run fast with Altivec and the PPC architecture, therefore, they are inefficent when run on an x86 processor

Photoshop is also optimized for mmx and sse (don't know about sse2). And as there are no standard C extensions to program mmx or sse code (unlike for altivec), I think it's safe to assume that those filters are even written in assembler.

Re:Misses the point

[binary paladin]

Pretty much.

I also like the way the config scripts run. I might seem like a small deal, but I'll tell you... if I could find something that used debs but had a startup system like Gentoo, I'd be really happy.

I love portage and frankly, I don't mind compiling my software, but that wasn't the allure of Gentoo. It's not even the configurability or customization factor. I find that it's just plain nicer to use. This almost seems opposed to what I find in a lot of its users.

ebuilds (like for the nVidia binary driver) just work.

Re:Misses the point

[Lord of the Wazz]

Also, the kernel compile is unfair, because gentoo-sources includes a whole load of patches that Mandrake and Debian don't.

From the article:

The same 2.4.21 source was copied to all machines and compiled using the same options. However, it should be noted that the Debian system used gcc 3.3.1 whilst the Mandrake and Gentoo installations used gcc 3.3.2 .

Which leads us to...

Finally, what's with measuring compile times? How is that a fair way of measuring performance?

I don't see any problem with comparing compile times. In the case of Mandrake and Gentoo, it's the same version of gcc being used and the exact same hardware being used each time so it's a reasonable comparison.

it looks like the guys doing the testing got their CFLAGS wrong

That may well be the case, but all that goes to show is that it's really easy to mess things up on Gentoo so that you get crap performance.

Re:Misses the point

[Anonymous Coward]

Great! Where can I get an upto date version of this distro though? I checked their mirrors, and they're all ancient.

Re:Misses the point

[VPN3000]

Yeah, the CPU is definitely a bottleneck when I try to run Doom3-alpha at 1600x1200.

Re:Misses the point

[Anonymous Coward]

What are the version/revisions of GCC on those machines?

RTFA

What hard-drive optimization did they set-up?

RTFA

What are the versions/minor revisions of the Gnome window manager on all those boxes? and GTK?

Are you seriously claiming that a combination of a specific version of GTK+ and Gnome caused Gnumeric to slow to the point that it took 8 minutes to open a spreadsheet? More to the point, what fucking difference does it make to the user? All they see is that their spreadsheet is fucked, and they don't want to hear you whine about toolkit versions. Thats a serious bug.

In other words, please shove your Gentoo up your Gentoo fanboy distended rectum. Thank you.

Re:Misses the point

[Alan Partridge]

"Hell, the default configruation of the Dell system with the 3Ghz P4 is 512MB of DDR - the high-end PowerMac (which they undoubtibly used) has 1GB."

What are you talking about? The default config for the dual 2Ghz PowerMac G5 comes with 2 x 256MB PC 3200 RAM.

Re:Misses the point

[BenjyD]

> Hmmm... that extra 200 mhz of FSB doesn't really do much,
> does it? It's still limited by memory bandwidth. Moreover, that
> bandwidth is shared between both processors in a DP system.

That extra 200Mhz means that when the memory bus is supplying the processor with data at maximum rate, the processor still has spare FSB bandwidth to talk to the hard discs, network card etc without slowing memory access. This happened with the early Athlons too.

I don't get this whole holy war thing.

Re:Misses the point

[buchanmilne]

I don't know how many rpms I download

Mistake number one. If you're going to use binary packages, use the tools your distro provides (in the case of Redhat you're a bit out of luck here) to install packages.

Downloading an arbitrary binary package from some arbitrary user is just as bad as using a binary build from someone else's Gentoo box.

And the vast majority of the time the packages doesn't need the new version, it just so happened that the developer was playing around with some CVS version of the library when he packaged the software.

It's not always that simple. It may be that a different package that requires the library *does* require the latest version of it. Anyway, if the developer took a CVS snapshot, it's his responsibility to test all the software that depends on it, and rebuild them if necessary.

It wasn't till recently that I realized gentoo's potential for easing the rpm-dependency hell that other distro's have.

You only have dependency hell if you don't use the tools provided by the distro. Many of the issues will affect source-based distros, but in most cases the source-based distros don't worry about it, since they're happy to have the user wait another day to rebuild all the dependant packages with different use flags to get one feature ...

Anyway, regarding Mandrake (which has pretty decent package management tools, and I am pretty sure you didn't give it a fair run), in cooker and cooker contrib, of the 4000 packages, there are only about 30 packages that have dependencies not satisfied at present (maintainers get spammed by a bot about this weekly), and many of them are either obsolete packages, or the dependencies *are* met, but the bot doesn't like them (file-based dependencies are not preferred, but do work although you get warnings from the bot).

Dependencies will *never* go away, if you choose not to use the tools that you are provided with to deal with them, that's your problem, not a distro's problem.

Re:Misses the point

[Anonymous Coward]

GNU/Debian/Linux All the way ! It is the most Free !

Gnu !

Re:Misses the point

[Anonymous Coward]

apt-get dist-upgrade
updates everything.

Re:Misses the point

[DrXym]

And it's binary patches which make it such a pain in the arse to keep Linux up to date. Really, everytime someone finds an exploit in glibc or the kernel which involves a few line patch, everyone is expected to download a 20Mb package!

That's fine if you happen to be broadband but sucks if you don't. In an ideal world, everyone would be motivated to waste the hours to grab the update, but how many bother, especially with the workings of Linux becoming more opaque and the users less knowledgable? The consequence is lot more unpatched Linux machines with all that entails.

Linux desperately needs binary patch support for this reason. I wonder how hard it might be to do. If someone has RH8 - i.e. a known configuration with RPMs of a certain version, it should be possible to produce incremental patches from that baseline that were a mere fraction of a full download. Instead of a 20Mb download, users would be faced with patches measuring in a few hundred k. Instead of hours to download, we're talking minutes at most. It even opens the possibilty to have an automatic download (and installation) functionality built into Linux, much like that found in MS Windows.

But MS Windows only uses automatic updates for small hotfixes for the same reason that service packs are humungus. How long before MS produce incremental patches for their OS? Here is the chance for Linux to take the lead for once.

Re:Misses the point

[Anonymous Coward]

You also forget that most other non-hacked CPUs have boatloads of GPRs(usually 20 or more, 32 on most powerpcs) whereas the x86 kludges have ~8 available. Also more recent x86s tend to have VERY deep pipelines which means that any stalls, branch mispredictions etc are very costly whereas powerpc and most other processors have relatively shallow pipelines meaning a smaller penalty from the same misqueues.

#2 I seriously doubt that photoshop is "more optimized" for powerpc any longer. Not much is really actually optimized for powerpc as the primary development is usually performed for x86 systems which are then hacked to just compile & run(more or less) on whatever architectures are supported. This is even true of open source, e.g. gcc v. Metrowerks(or even MrC on older OS versions).

Re:Misses the point

[Anonymous Coward]

I recently performed some tests with HP Itanium systems regarding kernel compile time. I tried on the same machine compiling a kernel where the source was on local disk and again where the source was in a ramdisk. I found that moving the source to ram disk (and thus effectivley having a disk with a few nanosecond latency and 3gig/sec transfer rate) made little differece, compile time dropped by about 20 seconds. This was using the -j3 option (2 cpu box). My conclusion was "its the cpu speed stupid".

Re:Misses the point

[CableModemSniper]

man xdelta.

I bet it wouldbn't even be too hard to automate, it even has a special case for compressed files. (ie you could make a diff against the rpm itself and reinstall as opposed to fiddling with "live" files.)

Re:Misses the point

[E_elven]

Because it's almost identical.

Re:Misses the point

[oxfletch]

But the cpu beign 90% utilized includes time for memory stalls ...

I've done a lot of testing on kernel compiles ... on machines with enough memory to cache the whole thing. Do a second compile afterwards with warm caches - it takes almost exactly the same time - it's NOT IO bound. (normally done with "make -j N" vmlinux) where N is 2x the number of CPUs.

Re:Misses the point

[lscotte]

Except in this case they all had the same hardware on each machine...

But different drivers. They used a mix of vesa framebuffer and SIS drivers, thereby invalidating the test as an apples-to-apples comparison. This test is not at all meaningful as a generalized performance test.

Re:Misses the point

[cperciva]

man xdelta.

No, please don't. xdelta is horrible for security patches -- the patches are much larger than necessary (eg, a factor of 5 larger).

BSDiff exists for a reason.

Re:Misses the point

[Anonymous Coward]

Here is the chance for Linux to take the lead for once.

Yeah! Take the lead, by copying what FreeBSD already does.

Re:Misses the point

[CableModemSniper]

Ok was just a thought.I was unaware of BSDiff.

Re:Misses the point

[dubiousdave]

You can't possibly be suggesting that Apple benchmarks should be looked at with suspicion. Steve would never mislead people.

Re:Misses the point

[QuintLeo]

For most of the work most of MY machines do, the bottleneck seems to be a COMBINATION of CPU and Memory limits - mprime/Prime95 is highly memory intensive, but I've seen significant improvements from kicking the CPU multiplier up on my Thoroughbreds WITHOUT kicking the FSB up.

I get even MORE improvement from kicking the FSB up *too*, though....

The rest of the work done by most of my machines is RC5 - and that is VERY CPU limited, as the core code fits easily into cache on anything from a Celeron up, and mostly so on the K5 and P5 generation CPUs.

Re:Misses the point

[ecchi_0]

I am not much of a Linux guy, but aren't those just display drivers that communicate with X? How does that effect compiling time/loading time?

Re:Misses the point

[Anonymous Coward]

du /usr/portage
Oooops... that optimizes for size for sure.

Re:Misses the point

[floamy]

There is no gcc 3.3.2. If he meant to say 3.2.3, which most Gentoo people are running, then we have a big problem. gcc 3.3.x not only compiles alot faster than 3.2.x, but it also often generates faster binaries, especially with sse/sse2.

Re:Misses the point

[nagora]

You're right, I don't, and neither does anyone else.

Yeah, I can see why that's "Insightful": no one uses the Gimp or has to analyse large datasets with their computer. What a bunch of fucking brain surgeons, as they say.

TWW

Re:Misses the point

[Anonymous Coward]

"It even opens the possibilty to have an automatic download (and installation) functionality built into Linux, much like that found in MS Windows."

Why the hell someone in his mind would want automatic updates!!!???

On the other hand, you ALREADY can do this with Linux.

I use Debian Sid (which, as it is unstable has a ton of daily updates: 57 today, for instance, due mainly to the new KDE version; this means almost 13MB downloaded).

Then I use cron-apt so the system automatically actualizes the packages list and download those to be new to my system. If I were so crazy to automatically install them, I would be able to do so. Of course, I want to see what's going to be updated, and I decide what I want updated and what I won't.

Still, yes, it would be a great advance being able to download only proper deltas instead of the whole package.

Re:Misses the point

[arkanes]

Maybe because it's true?

Large IO sets are much more bound by memory IO than CPU. Performing many calculations on a small set of data (so that the memory IO doesn't matter) is a pretty rare task.

(Oh, and there's no fucking point complaing about whatever a mod chose to do - I don't care about it, and don't control it, and the moderator who did it won't see it. Metamod if you care about crap like that)

Re:Misses the point

[DrXym]

Novices and consumers want automatic updates, since they expect their computers to just work. Binary patches offers an ideal chance for Linux to offer this. Otherwise Linux risks being polluted with exploitable boxes because people are too lazy (or busy doing other stuff) to update their machines and everyone suffers.

I don't suppose people who like to control their machines would like a feature, but for consumer level it's good enough. Personally I wouldn't mind the feature but I would prefer the OS asked me before installing them.

Re:Misses the point

[Anonymous Coward]

He was just pointing out that you're a retard. Although that should probably get moderated -1, Redundant instead of -1, Flamebait.

Re:Misses the point

[RzUpAnmsCwrds]

Hmmm... when you're running a Photoshop filter, you shouldn't be writing to the disk or the network. And if you are, you have bigger problems (swapping yourself to oblivion)

Re:Misses the point

[RzUpAnmsCwrds]

"You also forget that most other non-hacked CPUs have boatloads of GPRs(usually 20 or more, 32 on most powerpcs) whereas the x86 kludges have ~8 available."

Yes, the number of registers is an issue on x86. Luckily, Athlon 64 has greatly improved upon this (when it is running in 64 bit mode). Some apps don't really need the extra registers, some do - I don't know if Photoshop is one of those apps.

"Also more recent x86s tend to have VERY deep pipelines which means that any stalls, branch mispredictions etc are very costly"

Yes, the P4's 20 stage pipeline is quite long (Athlon is actually quite a bit shorter with 10 stages). However, the P4 has what is, by far, the best branch predictor in a production CPU. It also has various caches to reduce the penalty of a branch mispredict. Clock for clock, the P4 is not the fastest CPU - but it doesn't have to be. It only has to be 2/3 as good as a PowerPC CPU per clock.
"#2 I seriously doubt that photoshop is "more optimized" for powerpc any longer"

What I said was that the algorithms were designed in such a way that they ran fast on a PowerPC. Adobe is probably reluctant to change those algorithms because they might change the functionality of the filter.

Anyway, my point was that the statement that the G5 has much more memory bandwidth than the P4 is false. The P4 has equivelent memory bandwidth - 6.4 gigabytes a second.

Re:Misses the point

[RzUpAnmsCwrds]

"And as there are no standard C extensions to program mmx or sse code (unlike for altivec), I think it's safe to assume that those filters are even written in assembler."

No, but most good compilers (e.g. GCC) can optimize for SSE/SSE2. Also, ICC and other compilers have their own extentions for programming SSE/SSE2.

I know that Photoshop is optimized for MMX/SSE/SSE2. Heck, maybe they did a great job of it. My point was that the algorithms were developed to be efficent on the PowerPC, and that they just don't run as fast on a different architecture.

Re:Misses the point

[Anonymous Coward]

It's not offtopic at all. The article mentions gentoo sources and I was showing my experience with compiling the two kernels on the same system. Vanilla compiles much faster than Gentoo because there are so many patches for the gentoo sources. That could be the reason for the discrepancy. Maybe if the moderator RTFA it wouldn't be a problem. I guess it's just another anti-gentoo zealot given moderator access.

Re:Misses the point

[Xabraxas]

If you compile different drivers into the kernel then it will make a difference in the compile time of the kernel, which was one of the tests.

Re:Misses the point

[Halo1]

No, but most good compilers (e.g. GCC) can optimize for SSE/SSE2.

Most gains in graphics filters are gotten from vectorising the code. GCC cannot vectorize anything automatically afaik. The only think it does when you tell it to use sse, is replace all floating point operations with sse/sse2 equivalents. The reason is that they are more RISC like (as opposed to the stack-based 80x87 FPU), which results in GCC's optimizer being able to do a better job. ICC otoh is able to vectorize code on its own.

Also, ICC and other compilers have their own extentions for programming SSE/SSE2.

This locks you to a certain compiler however...

My point was that the algorithms were developed to be efficent on the PowerPC, and that they just don't run as fast on a different architecture.

Do you have more information about this? (I don't mean benchmark results, but background on how they did this)

LFS

[gibbdog]

LFS from scratch is my distro of choice :-)

Re:LFS

[Anonymous Coward]

obviously you like to spend more time tweaking and installing your system than actually using it to get stuff done.

each to his or her own of course, but you obviously arn't really using your computer to accomplish anything other than learning about an operating system.

Re:LFS

[Anonymous Coward]

LFS from scratch is my distro of choice :-)

And I have a girlfriend.

Re:LFS

[netsharc]

I tried LFS, but got confused after a while, how do you track what you've installed and into which directory?

I feel Gentoo is one-better than LFS, it does the same thing (compile from scratch) but with a useful package manager. LFS on the other hand is good to learn how a Linux system works, although one could argue reading a HOWTO about it would be faster than waiting for the compile to complete.

Re:LFS

[Anonymous Coward]

And I have two. Of course I use windows.

Hmmm wonder how many I could get if I just gave up on computers?

Re:LFS

[Anonymous Coward]

its LFS not LFS from scratch
see the from scratch is right in there so you dont need to say it again. I bet you say PIN number too/

Interesting, but I have to wonder.

[Enahs]

I have to wonder what the reviewers consider to be a "default" install. For example, did the reviewers remember to build in support for their IDE controller (if that's what they use)? If so, is DMA enabled for the Gentoo box, and is it for the others? What kernel did they use? Did they use gentoo-sources or did they use another?

Maybe to the uninitiated this seems informative, but to me it doesn't.

Re:Interesting, but I have to wonder.

[Anonymous Coward]

>> Maybe to the uninitiated this seems informative, but to me it doesn't.

That's because you lack experience to read between the lines.

Have patience and in some 10 years you will be able to make much of even a short, beautifully written article like that.

You'll also become less arrogant, which may help you deal with people.

Have a nice day. And try to learn, this time.

Re:Interesting, but I have to wonder.

[Anonymous Coward]

Nice troll, very nice. One of the best trolls I've seen here in weeks. Thank you.

Re:Interesting, but I have to wonder.

[Enahs]

It'd be a troll if I didn't mean it. :D

I've got RH9 on this machine right now; I happened to be checking my email before moving a chroot'ed Gentoo build over to another drive for installation when I read this article.

Gentoo may take massive amounts of CPU time to build, but it's worth it, oh yes, it's worth it, IMHO.

Re:Interesting, but I have to wonder.

[BrookHarty]

My own experience has been that Gentoo outperforms Debian on my hardware, but only after I've done some tweaking on Gentoo. YMMV.

How true, I wish we had a 3dmark type program for Linux, where we could test X performance in 2d/3d, audio, hd, cpu, mem and even latency for each area. Maybe even report performance and features for OpenGL, to see what the drivers do and dont support.

A good benchmark program could be used to see if newer kernels are really faster (ie 2.6) or even those nice pre-emptive kernels patches.

Thats one part lacking in Linux/unix, hardware testing/benchmark programs. Not counting FS benchmarks, there are handfuls on freshmeat.

Thou, maybe a benchmark program would make Linux look slower to windows in the desktop area. (Any idea?)

Re:Interesting, but I have to wonder.

[Anonymous Coward]

A good benchmark program could be used to see if newer kernels are really faster (ie 2.6) or even those nice pre-emptive kernels patches.

http://members.optusnet.com.au/ckolivas/contest/

Try doing some research.

Re:Interesting, but I have to wonder.

[Anonymous Coward]

http://members.optusnet.com.au/ckolivas/contest/

Ahh, see AC, you didnt really understand did post did you.

That program only benchmarks kernel responsivness. Its not an Entire benchmark program like 3dmark or Sandra sci-soft.

try reading posts

Re:Interesting, but I have to wonder.

[bogolisk]

My own experience, has been than Debian outperforms Debian on my hardware, but only after I've done some tweaking on Debian (the latter)! :-)

In theory, there is no difference between theory and practice. In practice, there is. -- Yogi Berra

Re:Interesting, but I have to wonder.

[arkanes]

Thats exactly the point, as some other people in this thread have mentioned - there's alot more to getting good performance out something that just passing -arch=pentium4 -O3 to your kernel. Of course they didn't go seeking out special gentoo kernels. They used the stock one IN THE INSTRUCTIONS. They didn't patch the Mandrake or Debian kernels, either (Mandrakes stock kernel doesn't not have the interactivity patches). Gentoo, when used in the normal case (ie, someone who follows the instructions), will not have noticably better performance for most tasks. So there you go.

Any gentoo baiting you seem to read from the article is certainly in your perceptions. There no "bias" against gentoo. They don't jump up and down yelling about how much gentoo sucks.

Re:Interesting, but I have to wonder.

[Anonymous Coward]

Sounds like a great idea for a program! When are you doing to do it?

Slow?

[peripatetic_bum]

Can this be correct. Debian turns out to he fastest?

Anyway, I like the idea of gentoo, and I saw I a lot of Debian users head over to gentoo because the idea of controlling everything including the build was nice, however, I saw the gentoo idea also pretty much die, since a log of these users are power desktop users and not everyone could wait 3 days for X to build.

What I like about debians packages is that if you do make a mistake you can always pretty much correct the package by fixing your souce list or goingt o packages.debian.org and getting the older working package and installing it manuaall with a simple dpkg -i old_package.deb.

In gentoo, you had to rbuild to the whole thing, whihc with x coud take forever. And so what I saw gentoo suddenly doing was having a lot of pre-complied binaries start being provided by gentoo because they saw the problem with building taking forever, and so it kind of killed the whole idea of building for yerself, in which case, if you are going to stick with built pacakged why not have them maintained by some of the best developers around (ie debian)

The othjer thing I noticed is that a lot of developers of software acutally use debian. I've noticed many a time that some cool software wa being made and the developers wouls provide source and they would provide a debian package and nothting else. Ie Debian appears to be the preferred developer's distro. In this I would like to hear discussion,.

Thansk all

Re:Slow?

[dJCL]

What are you people running on? My compiles take about 3 days for X sure, but I am using a 500mhz system? WTF is up... ah well Debian rocks.

Re:Slow?

[Anonymous Coward]

You've never built X before have you. Didn't take long at all, mozilla took longer.

Re:Slow?

[Anonymous Coward]

3 days? I'm using a 667MHZ and it only took a few hours. Works great!!! Even better than my RH8 box.

Re:Slow?

[antiMStroll]

....these users are power desktop users and not everyone could wait 3 days for X to build.

Desktop power users on 386's are a rare breed nowadays. I don't know how long it took X to build on my P2 366 w/ 192 meg RAM because I started it before going to bed and it was done in the morning. Maybe these power users should consider hardware before choosing distros.

And so what I saw gentoo suddenly doing was having a lot of pre-complied binaries ....

Lots? OpenOffice has a precompiled option, Opera does because it isn't open-source, but which other packages do you mean? To the best of my knowledge almost all Gentoo packages are still source-based and more than a few have new options to pull the newest source directly from the CVS tree. Gentoo is expanding into more source-based options, not pre-compiles.

None of this is to say that Debian isn't a better developer's platform, or faster than Gentoo. I wouldn't know, but your post is very misleading.

Re:Slow?

[ctr2sprt]

Yeah, I was a Debian user and tried Gentoo when it first came out. I'd used FreeBSD, so I knew and loved the ports library, so I was excited about Gentoo. Unfortunately, the initial releases seemed broken in several severe ways. Half the software in Portage wouldn't compile at all, and I didn't really feel like digging into the source to find out why. I'm not some idiot newbie, I'm a computer programmer and have been using Unix for nearly 10 years. I just wanted the install to work, and it wouldn't, no matter what I tried.

Anyway, obviously Gentoo has improved since then, but this is a concern of their Portage system (developers accidentally breaking parts of it, and if those parts happen to be gcc, look out!). It happens in FreeBSD every now and then, but it's not as big a deal there since the BSDs use actual releases: the ports are all tested against a specific release and verified to compile, then the ports library is frozen until the next release comes out (which will be tested similarly). So here's a question buried in this rambling anecdote: does Gentoo provide a way of getting "stable" ports, or is the entire OS like the "unstable" branch of Debian?

Re:Slow?

[XO]

Three days for X to build? only took about 8 hours on my P3/600 with 128MB ram, and about 10 hrs on my AMD K6-2/450 with 288MB...

Re:Slow?

[XO]

I just compiled X 4.3.0 on my Debian 'unstable' system, and it definitely did not take even 10 hours.. i left home, came back a bit around 10 hrs later, and it was all installed.

Re:Slow?

[1lus10n]

not sure if your just trolling or not but anyways:

the gentoo "idea" is far from dead my friend it is STILL the fastest growing distro, and is still #3 overall in world wide usage (behind redhat and Suse)

oh and just for you information you can use emerge to get pre-compiled binaries that were customer built for your arch. such as athlon.

as far as cleaning "mistakes" out of gentoo, well thats really hard, like so:

emerge -C someebuild removes
emerge someebuild.2 installs
OR
emerge -K somebinaryebuild.2 install the binary.

my head hurts from doing all that work.

Re:Slow?

[peripatetic_bum]

Sorry, My post is more confusing than mis-leading. :)
I had two things to say:

1. relevant to the topic: Has gentoo become just another distro in that it provides lots of binaries thus ignorning its advantage

2. Not releavatnt to topic; is debian the developer's distro of choice?

G

Re:Slow?

[XO]

Oh, I need to mention.. PIII/600 MHz 128MB ram..

Re:Slow?

[peripatetic_bum]

8 hours? 10 hours?

Are the advantages that great?
I am very curious again

Thanks

Re:Slow?

[Rhone]

Speaking as someone who used Debian ("unstable") for years and switched to Gentoo a few months ago....

First of all, Gentoo has definitely improved since you tried it. It has worked quite well for me, and I can't recall anything failing to compile. I imagine that if something as significant as gcc were broken, the problem would probably be rectified rather quickly. I worry more about fringe programs that few people are using.

Anyway, my answer to your question of whether Gentoo is like the "unstable" branch of Debian would have to be "yes and no".

It's like the unstable branch in the sense that the newest versions of software are easily available to you soon after they are released. It's also like the unstable branch in the sense that you can try out that new software before it has undergone weeks or months of testing with the rest of the system, and like Debian unstable it is best for the geekier crowd who can deal with unexpected problems here and there (I'm being somewhat hypothetical here, because I haven't really had any such problems in Gentoo).

It's also like Debian unstable in the sense that it's actually usually more stable than the released versions of many of the rpm-based distros. (Though it's been a while since I last used Mandrake or RedHat, so perhaps they have improved their stability as well?)

It's unlike Debian unstable in the sense that you have more power to fix the problems that may come up (at least problems that can be blamed on the distro and not on the software itself). I felt kind of helpless with Debian unstable when dpkg would choke on a deb. You can't do much (that I know of) to edit an already packaged deb, and getting the src deb to try to find the problem seems like kind of a daunting task for a non-Debian-developer.

In Gentoo, if there's a problem in an ebuild and you don't want to wait for it to get fixed... well, you have all of the ebuilds right there on your hard drive under /usr/portage. Ebuilds are fairly straight-forward; it doesn't take too long to learn how they work so that you can make simple edits, or even create your own. Also, ebuilds for old versions usually hang around for a while, so if a program is working fine for you and then there's a bug in a new version, you can downgrade easily.

One of the features I like is that you can put your own ebuilds in /usr/local/portage, and they will be handled gracefully by emerge. And /usr/local/portage takes precedence over the stuff in /usr/portage, so you can use it to replace existing ebuilds in a way that won't be overwritten next time you "emerge sync".

In Gentoo you are really never helpless.

Re:Slow?

[lewp]

You very well may know this already, but Daniel Stone has an unofficial apt repository with XFree86 4.3.0 packages. I've been using them for months with great success.

http://penguinppc.org/~daniels/xfree86/README

If you just like compiling X, or have some other reason to, forget I said anything :).

Re:Slow?

[Micah]

I'm a former Red Hat user who switched to Gentoo a month and a half ago. Basically, things just work now. Whenever I've had an error with an emerge (a compile error or some other kind of emerge error) I've just searched the Gentoo forums and always found a solution.

Re:Slow?

[lewp]

FWIW, modifying and recompiling a Debian package isn't as difficult as you think it'd be. No more difficult than editing an RPM .spec file, IMHO.

Maybe Gentoo ebuilds are easier than that (I don't know, I've never read one), but considering you probably just need to fix a typo or make a simple change it's not likely to make too big a difference.

Re:Slow?

[jyristys]

So here's a question buried in this rambling anecdote: does Gentoo provide a way of getting "stable" ports, or is the entire OS like the "unstable" branch of Debian?

Kind of, you can choose between stable or unstable packages, I guess Gentoo stable has similar versions of packages as Debian unstable.

The reason I'm using Gentoo right now instead of Debian is basically that I wanted kde 3.x:
root@debian # apt-get install kde
blahblahblah.. Could not install this and that, sorry, broken packages.
root@gentoo # emerge kde
blahblahblah, here you go, enjoy your nice new kde.

Re:Slow?

[mcp33p4n75]

Yes. Whether or not you have stable packages depends on your ACCEPT_KEYWORDS setting in /etc/make.conf. Say, you want stable on x86, you do:

ACCEPT_KEYWORDS="x86"

Or, you want unstable packages (like me :):

ACCEPT_KEYWORDS="~x86"

It's pretty simple, really.
As some worthless anecdotal evidence, I have never had a problem with the stable packages, and i've been using Gentoo for nearly a year. As for the unstable, most of the problems I have had are build/configuration errors.

Re:Slow?

[netsharc]

Indeed it takes a while to compile things. I'm reminded of a time when a misconfigured MAKEFILE (I assumed that's what it was) deleted the main qt library. I was in KDE, boom all of a sudden it went black and I was back at the console. So, re-emerged Qt. This time I made a backup copy of the file though, and sure enough I needed it a few (minutes/hours) later as I fubar'ed an ln -s and zeroed the qt lib file.

Re:Slow?

[PugMajere]

You can only configure it based upon keywords?

Debian (well, apt really) has a "Package Priority" system, that lets you selectively pick distributions, versions, etc, of packages to install.

The whole explanation gets complicated - but I have a machine at work that has Perl 5.6.1 installed, but the rest of the distribution comes from either Debian/testing or Debian/unstable, where Perl 5.8.1 is standard. (I have binary compatibility problems that require me to continue doing development against Perl 5.6.1)

Is that easy to do in Gentoo? It seems that a keyword based system would be a bit... limiting in that respect.

(Keep something in mind - most of computing revolves around, "is it good enough?" and not, "This is the best." Gentoo may be neat, but Debian is good enough for me, and the benefits that Gentoo may or may not provide don't seem big enough to make up for the time spent compiling everything, even in a setup where I could use distcc across somewhere around 10 machines.)

Re:Slow?

[rendler]

A .deb is just an ar archive, you could have unpacked it and looked through all the pre/post install scripts (if any) or any of the files in there. Fix the problem and pack it back up. Definately not the the newbies out there, but it's at least another option ;)

Re:Slow?

[mapMonkey]

Certain packages are "slotted", such as gtk and gtk2, so that you can have both versions installed on the same machine. Perl is not one of them AFAIK, so having 5.8 and 5.6 on the same machine would require some intervention on your part.

Re:Slow?

[Anonymous Coward]

X takes me about 2 hours, and that's using some pretty heavy optimizations, however I'm using a Athlon-XP 2500+(Hey for $90 retail I think it wasn't a bad choice), using ccache 4.4 will only take about 30 minutes. If compiling X is a problem, then use Knoppix, so you still have functionality and you can build your system.

Re:Slow?

[Anonymous Coward]

No Gentoo has not provided a lot binaries. There are binaries for packages like KDE, Gnome, OpenOffice, X, and some other programs on the 1.4 CDs. This is for people with slower machines, or don't have time to build them, or they will do an emerge -e world later.

Re:Slow?

[Anonymous Coward]

Xfree86 is such a bloated fucking pig. It used to take 8 hours to build X11R5 on a Sun 3/260 (25MHz 68020) with 32MB RAM. Took 4 hours on a SPARCstation IPC (25MHz SPARC v7) with 16MB. Both had 3600RPM SCSI disks. Neither needed Perl to generate the fucking fonts for it.

Re:Slow?

[jyristys]

8 hours? 10 hours? Are the advantages that great?

Those times were with machines several years old, on my spanking new xp2500+ 512Mb 'emerge kde mozilla gimp kdevelop jedit' (+ others, you get the idea) took less time than it took me to sleep, go to work, work for ~8 hours and come back home.
I don't really need any advantages to justify leaving the machine compiling when I'm not using it.

Re:Slow?

[XO]

I've heard that there are some pretty huge advantages in compiling X for your specific machine.. and I needed X4.3.0 for the PCI Radeon support, which did not come in any Debian installation.. so I just let it do it. :)

Re:Slow?

[dubiousdave]

You can be more selective on which unstable packages to install. For one thing, you can pass the ACCEPT_KEYWORD= environment variable when you issue the emerge command. The environment variable takes precedence over the config file. I've done this to install a few packages that were still marked as unstable, such as MythTV v.0.10.

I believe you can also configure it to specifically use unstable versions of some packages in the config files, but I don't recall how. I'm pretty new to Gentoo myself, and I haven't had enough need to bother learning that trick.

BTW, even though I am far from a Gentoo optimization guru, I have been amazed at the speed difference between my Gentoo Stage 1 install versus my previous RedHat 7, 8 and 9 installs on the same machine. It took me a while to figure out why my first kernel build (after installation) failed after only a few minutes. It didn't. It was just finished. It made the previous 24 hours I'd spent doing the Stage 1 install worth it, I think.

FreeBSD's ports

[Anonymous Coward]

I don't use Gentoo (When I use Linux, I use Slackware), but I do use FreeBSD and its ports collection.

Purported performance gains are one thing source packages give you (although I don't enable super optimizations because you never know when gcc bugs with -march=pentium4 -O3 or whatever will bite you).

There are two major reasons I like installing from source, though. One is that you can customize the build to your system; lots of software packages have various compile time options, and when I have the source I can choose exactly how it's going to be built.

Another thing is that when you install from source, you can hack the program to your heart's content. On my desktop box there are around 15 programs that I have to modify to get to act like I want (from simple things like getting cdparanoia to bomb immediately when it detects a scratch to halfway complex things like rewriting parts of klipper and XScreenSaver, which now picks a random screen saver on MMB and lets me scroll through all screensavers with the wheel =).

I don't modify stuff on my servers, but I still get to choose exactly how things are built, which I very much enjoy.

Re:FreeBSD's ports

[Anonymous Coward]

pffft, whatever...

apt-get source

Change code to your heart's content.

dpkg --build

Pretty freaking easy, eh? Plus I don't have to waste time compiling anything I don't want to.

There is no Gentoo

Re:FreeBSD's ports

[Anonymous Coward]

Neither do I. Ever hear of FreeBSD packages? Of course not, because you're a debian zealot.

I would run FreeBSD waaaay before I'd run debian.

Re:FreeBSD's ports

[XO]

can I have those patches to Xscreensaver???

Re:FreeBSD's ports

[Anonymous Coward]

If you've got somewhere (FTP, whatever) I can send the patch, sure. Slashdot thinks it has too many junk characters.

Re:FreeBSD's ports

[hacker]

(from simple things like getting cdparanoia to bomb immediately when it detects a scratch to halfway complex things like rewriting parts of klipper and XScreenSaver, which now picks a random screen saver on MMB and lets me scroll through all screensavers with the wheel =).

I presume all of these patches were sent upstream to the maintainers of said packages, so they can evaluate them and roll them into their core versions, riiiiiight? Even submitting them to the project mailing list would be advantageous for those other users who might find them useful.

Re:FreeBSD's ports

[Anonymous Coward]

Generally not worth it. Most of the time these patches would alter behavior enough that they'd need to be compile time/runtime configurable, and I don't care enough to get the patches clean. JWZ's FAQ addresses the MMB issue, if you were wondering about that. But as I said in another message, I'll put the diff on an FTP somewhere if space is provided.

Bugfixes I send, personal changes I tend not to. C'est la vie.

Re:FreeBSD's ports

[Anonymous Coward]

NO DOUBT.
Everyone knows that debian is for little bitches.
No wonder Gentoo has more users than debian after such a short time.

Re:FreeBSD's ports

[XO]

The email address is right there ^^ :)

Remember the KDE mandrake/gentoo fiasco?

[BrookHarty]

While the posts are starting, and people are saying Mandrake would never be faster. Lets go back earlier this year...

Remember the KDE optimizations that where not included in the Gentoo source release? Everyone was wondering why KDE was faster on Mandrake. There where talk for over 2 months before people realized it was an option Mandrake was compiling with.

Myself, Gentoo's biggest feature was the kernal compile options, adding patches for pre-emptive mulitasking, and improved responsiveness. I noticed the improvements on all my machines, but the compile times where a draw back. And sometimes the applications wouldnt compile.

Mandrake while my favorite choice, doesnt include the best pre-emptive kernels. Which do make a noticable difference. So after installing mandrake, and putting a newer kernel on the system normally takes care of that.

I'm just waiting till beta2 of mandrake cooker 9.2 with the 2.6 kernels, that should make Gentoo and Mandrake on par for speed.

From a LFS perspective

[shoppa]

I don't use Gentoo, but I do use Linux From Scratch, and I do see substantial improvements with command-line type activities: A kernel build on a Athlon is about 20 percent faster when I do it with a custom LFS build vs a stock RedHat installation.

Most of the comparisons in the article were for X-related graphics applications, and while they were comparing the versions of the applications, they were not comparing the libraries underneath them (glibc, X11, and probably the window manager too come into play) and they should've compared versions there too. It becomes complicated because for a typical X11-based app there are probably several dozen libraries involved (in addition to all the configure-time options for them...)

Re:From a LFS perspective

[jay-be-em]

Unfortunately not all of us have oodles of time on our hands.. This is where debian comes in.

Re:From a LFS perspective

[Anonymous Coward]

I don' have oodles of time either, but my cron job does :)

Re:From a LFS perspective

[Sloppy]

Unfortunately not all of us have oodles of time on our hands.

Me neither, but my computer sure does. If it wants to compile stuff when I'm not around or when I'm asleep, that's fine with me.

Why use it?

[Realistic_Dragon]

I picked Gentoo because it was Free and free, and because emerge has IME one big advantage over APT - one well updated, consistant, all encompasing, repositry.

OTOH my laptop runs RedHat, because I needed at least one machine running it to stay current with where they dump configs (it's the distro they use at work). Coupled with Apt-RPM it's competent enough, and I have no major problems with the performance.

So yeah, I have to agree with the article - you may like it one way, others may want to do theit own thing. No matter what you chose, you (probably) have binary compatibility, so who gives a sh!t about the holy wars, just as long as you aren't running Windows :D

Re:Why use it?

[Anonymous Coward]

The biggest problem with APT-rpm that I have is that APT isn't very "apt" when it comes to installing the new redhat kernels. I prefer yum (http://linux.duke.edu/projects/yum/). I haven't used apt in a while, but I got tired of typing rpm --rebuilddb after 75 updates I installed on my system. I don't know if they ever fixed that bug.
After running yum for most of a year, I definitely prefer yum to apt-rpm.

That said, I find it interesting that there's really no reason an APT-rpm or Yum can't be modified to do essentially the same thing that emerge does with Gentoo.

I've had my share of source RPM's that wouldn't compile correctly without some modification to the spec file. But I would like to see someone modify yum to use the source rpms, install all the devel packages and then install it. As an added benefit red hat, mandrake would make better source rpms explicitly for that purpose.

Gentoo similiar to *BSD?

[jtotheh]

I have never tried Gentoo but I ran FreeBSD for a while. With FreeBSD you have source for the whole system as well as for any "ports" you install. There are procedures for doing a "make world" that recompiles all of it. You can get the source changes to go to the next version and with a bit of chicken and egg stuff about compilers if that has changed, you can compile yourself the upgrade.

I ended up bagging it because there is a fair amount of stuff for Linux that is missing in BSDs (or I wasn't willing to expend the effort to get to run through compatibility mode). Java, Flash, etc. no flames please, I know some people work these things out - I just got tired of the hassles with it when I could rpm/apt-get it with Linux.

I thought the FreeBSD was really high quality though.

Is Gentoo a similiar model? Has someone used both?

Re:Gentoo similiar to *BSD?

[jtotheh]

oh yeah....after a while I got tired of all-day compiles.

Re:Gentoo similiar to *BSD?

[Gramb0]

I've used both Gentoo and FreeBSD and yes, there are similarities between the ports system used in FreeBSD and the portage system in Gentoo, although the actual way they are run is fairly different. For example in Gentoo you would install gaim by typing 'emerge gaim' compared to 'cd /usr/ports/net/gaim' then 'make install clean' in FreeBSD. Gentoo did have some nice ideas, such as USE flags, where you could define flags that would effect build options on any relevant programs, for example if you had the use flag +mysql defined then any programs that had optional mysql components in them would be compiled with them. Whereas -mysql would ensure that any programs with optional mysql support would definetly not be compiled with mysql support. Another nice feature of ports is package masking where packages that are reported to be unstable on certain architectures wouldnt install on that architecture (although you could override this if you didnt care about the potential instability) It was also fairly simple to upgrade programs in gentoo (could be just as easy in FreeBSD too, but i've not been using it for long so cant be too sure) since all you had to do was run 'emerge sync' to update the portage tree and then 'emerge -u world' to update all the installed programs (or you could specify individual programs instead of world) Anyways, thats enough talk of portage from me :)

Re:Gentoo similiar to *BSD?

[Gramb0]

oops, being a bit new to slashdot, i forgot to change the format method so my post is a lot harder to read, oh well i wont be making that mistake again :)

Re:Gentoo similiar to *BSD?

[marcovje]

Under FreeBSD this conditional including packages is usually named USE=1 or NO=1 for simple cases, or adding a separate -no port. (e.g. in no Gnome/KDE/X11 cases)

You can add these to your /etc/make.conf

I've used Gentoo and FreeBSD. While I Gentoo was nice, I stuck with FreeBSD, however this is more because of the quality of the release engineering done on the packages than the tools to install them.

Re:Gentoo similiar to *BSD?

[RdsArts]

And thus, pkg_add was born.

Re:Gentoo similiar to *BSD?

[tigga]

For example in Gentoo you would install gaim by typing 'emerge gaim' compared to 'cd /usr/ports/net/gaim' then 'make install clean' in FreeBSD.

In FreeBSD I'd rather type
pkg_add -r gaim
to install it. I do care about my time. It's kind of insane wait for computer to compile things if there are already available packages.

Re:Gentoo similiar to *BSD?

[Daengbo]

My problem after using Gentoo for about 5 months was emerge -u world to find I had all kind of broken stuff on my machine. Gentoo would recompile a library or mysql and all the programs that depended on it would start acting funny or randomly crashing. I got tired of recompiling my machine from scratch and went back to a binary distro which updates constantly, giving me much of the benefits that I had with Gentoo, but none of the headaches.

The Mindcraft method, against itself

[lavalyn]

Besides, before doing any comparisons on Debian vs. Gentoo they should have compared Gentoo vs. Gentoo on different optimizations. Like using -O2, -Osize, -mfp-math=sse. Comparing video drivers. Trying different filesystem types. And a whole gaggle of other configurables at compile-time.

You'd be yelling bloody murder if Microsoft sponsored a study without doing this sort of research before pitting Windows vs. Linux.

Re:The Mindcraft method, against itself

[Nucleon500]

I wholeheartedly agree. First, there needs to be a testing process as you describe to find what works fastest. When I started using Gentoo, the hardest part was chosing my CFLAGS. I have a P4, but there were reports that GCC -march=pentium4 was badly broken.

And second, the GCC people need to really pay attention to the results, and work that into the -march and -mcpu defaults, and fully test and document the real-world (as in compiling and using a whole distro) effects of this. I'm tired of people saying Intel's compiler makes faster code -- if that's the case, we need to do something about it. (Also, Intel's compiler can build the kernel, I wonder if it can be made to build Gentoo.)

Re:The Mindcraft method, against itself

[agurkan]

You'd be yelling bloody murder if Microsoft sponsored a study without doing this sort of research before pitting Windows vs. Linux.
What??? Microsoft sells Windows, if they make a comparison, of course they need to be a lot more careful. Plus, what's with the "bloddy"? What are we? British?
PS: No offense to British folks ;-)

Re:The Mindcraft method, against itself

[Arker]

The odd thing is that, from what I've read, a lot of Gentoo folk seem to be trying to compile everything with -O3. This is, frankly, bloody stupid. This turns on a lot of 'optimisations' that are only useful on a few programs and actually harmful for most, and is probably one of the reasons it looked bad in this test.

O1 is the safe level of optimisation. Even O2 runs the risk of doing more harm than good, although it's a fairly low risk. O3 runs a very high risk of doing more harm than good. In many cases Osize is probably the best option anyway, because I/O is more commonly the bottleneck than cpu capability.

And the processor optimisations can also be risky. Every processor out there is designed to run commercial i386 code as fast as possible anyway.

The big wins in compiling yourself are control of configure options, not compiler optimisations. I think source-based distributions are a great idea, but I have to wonder if most people using them right now are getting the benefits.

Re:The Mindcraft method, against itself

[garver]

On the other hand, the article does show that using gentoo does not mean you can assume better performance, which I think is the common perception. If you use gentoo for performance reasons, you should be prepared to tweak, rebuild, test, repeat.

It's silly when people use gentoo and simply build the way someone told them too. How is this different from a binary distro? Except for the hours of compiling, of course.

Re:The Mindcraft method, against itself

[hayden]

Besides, before doing any comparisons on Debian vs. Gentoo they should have compared Gentoo vs. Gentoo on different optimizations

And how many Gentoo users have done this?

If Mandrake had funded this study then your way would have the correct way. This on the other hand was a limited study done with limited time (by the sounds of the article). The testing was done with some fairly standard optimisations to see if it's worth testing further and appears as it is.

Re:The Mindcraft method, against itself

[j3110]

-Os is stupid...

Just try that on a P4.

If they really wanted to test performance, they would have used bzip2 -9. I don't even understand what they are testing. Are they testing how long it takes to start an app??? That can depend on where on the disk the files are stored. What a silly test.

Everyone go do bzcat time bash -c "linux-2.6.0-test2.tar.bz2 | bzip2 -9 >> /dev/null"... That will test how fast your system is more reliably. Two processes test the schedular a little, test pipes (IPC), test disk IO latency (bzcat reading data from the file), but mostly tests the raw speed of the CPU/optimization of libbzip.

Re:The Mindcraft method, against itself

[Arker]

-Os is stupid...
Just try that on a P4.

What exactly are you trying to say?

-Os decreases cache misses, and that's just as important on a P4 as on any other CPU.

If they really wanted to test performance, they would have used bzip2 -9. I don't even understand what they are testing. Are they testing how long it takes to start an app??? That can depend on where on the disk the files are stored. What a silly test.

Opening applications is something most users probably do a lot more often than running bzip.

Re:The Mindcraft method, against itself

[j3110]

Cache misses are far less important than branch prediction on many machines (especially the P4). It's true that more Cache will help performance, but I think it's quite fishy that faster memory doesn't seem to help very much. I can't tell a difference when my machine running ram at PC2100 (166 x 2) and PC2700, which I think is pretty fishy. In fact, I hardly noticed a difference from SDR to DDR. The chances are that you are only going to save about 5% of cache misses with -Os, and it's going to set you back at most 10ns, but probably on average one read that will take about 3ns, or about 13 clocks at 3Ghz with PC2100 RAM. With a 20+ stage pipeline, everytime a condition doesn't do what the processor expects, you're looking at about the same thing. There are many more conditions that go awry than pages of code and data that don't get loaded. It's easy to see what pages are going to be needed and can be prefetched by just looking at the next few instructions... cache misses really don't happen any more because the fetching code gets put in the pipeline in such a way that unless you are accessing random pages of memory, and suffer a RAS-CAS delay on every read, you are going to have the data loaded in Level-2 in the off chance that you will need it. A memory clock is a terrible thing to waste.

PC2700 lets you read at 2.7G/s. If you were just executing nop after nop, you could run at 2.7Ghz without a single cache miss. It all comes down to branch prediction.

Once you see that branch prediction is the problem, and that the RAM is more than adequately fast for everything but texture mapping graphics, the only thing that should concern you is making it easy on the branch prediction hardware. -Os does away with a lot of the optimizations that try to do this. Say I have:
y=0;
for(x=0;x10;x++) y=y+x;

You're size optimized version would be:
xor bx,bx
xor ax,ax
loopy:
add bx,ax
inc ax
cmp ax,10
jl loopy
some more instructions

unrolling loops I get:
xor bx,bx
xor ax,ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
inc ax
add bx,ax
some more instructions

You are gauranteed at least one branch miss, I'm gauranteed 0 branch misses.

Those of you that are very clever out there will be saying, "but if you loop smarter, you won't have that problem" and you're probably right. Here's the example best of both worlds..

xor bx,bx
mov cx,9
loopy:
add bx,cx
loop loopy

This is why hand optimized assembler will always be fastest amongst those that know the architecture well. Some smarter compilers may do the following:

xor bx,bx
mov cx,10
xor ax,ax
loopy:
add bx,ax
inc ax
loop loopy

This basically maps the iterator variable into 2 constructs so that you can read the iterator inside the loop (everyone does) and you count up (this is slower than counting down because of the loop instruction). If you just counted down, the 3rd example would be the fastest code.

Really good optimizers will say "there are no variables in this code, let me run it for you". Basically you end up with:
mov [y],45

The point of this exercise is, that by using the loop construct, the CPU's branch prediction should be accurate enough to do it your way, but it limits you to counting down. Any incrementing loop, you might as well tack on the length of the pipe*2 instructions and assume this as the total run length of the loop.

The other point is that when the processor sees mov [ES:BX],5 it's not going to be a cache miss because it's writing. When it sees mov AX, [ES:BX], it's probably not going to be a cache miss because it's going to load the whole ES as soon as it sees the program counter get anywhere close to it.

Cache misses come mostly from interrupts. When the schedular pushall's to the stack, then schedules another process, and does a far jump, that's where you're going to have cache misses just because the CPU can't see that much memory IO coming (you never know when an interrupt is going to happen).

Re:The Mindcraft method, against itself

[Arker]

You raise some good points. But the weakness of the analysis remains - you're analysing how this will affect a particular chunk of code. There's no doubt that -O3 and the architecture flags will do a lot of good on certain segments of code. That's why they're there. Use them on the right chunks and you can do very well.

But we're talking about using them, not just on this piece or that piece that will benefit from them, but globally. And there are plenty of chunks of code where they will hurt as well.

Basically we're both engaging in useless speculation to a degree here - neither one of us (I presume if you had you would have said something) has sat down and built systems, tested them, rebuilt them, retested them, enough to generate the hard data that would allow us to say for sure what's optimal here even on a single hardware setup. And that's what you'd have to do to be certain. And even then, the tests you chose to run would be an important part of the equation as well.

Because -O3 is going to help some code but hurt other, and because -Os is always going to make code more likely to fit in cache, I have a hunch that it's probably more efficient in the context of a global flag used in building the entire system. I have another hunch that -O2 is probably better than -O3 in that case as well, and that posing -O2 vs -Os the results would be very hardware dependent. But it's just hunches. Your points are quite valid, and I don't see any way we could really settle it except by agreeing on fair tests (a rather difficult problem in and of itself) and running them under both conditions, preferably on a wide range of hardware.

Do you?

Re:The Mindcraft method, against itself

[j3110]

Well, the best thing that could happen is that the GCC people test -O3 and -Os on different sections of code for each architecture, then add them to the -march, but I thought that's what they already did. -march=pentium4 will give you some of -O3 (-funroll-loops I think). I'm not certain on this though. I've not tested the actual assembler put out from the code, nor have I done the math to see if loop is faster than inc cmp and jmp more than just the clocks of the three instructions being reduced to one. It just only seems logical that the branch prediction would consider loop because you know if it's going to jump since it only depends on a register (can be passed back through the pipe) and you pretty much know you are going to not branch back if cx is 1. I can't say that I know that there are compilers out there that use loop as well, but I think it would be silly not to.

I think the best arguement would have to be made from someone working at Intel on their compiler (only for Pentium processors though) or someone on the GCC team that knows more about how the compiler works.

I also think that debian and mandrake packages are modified from the originals, so you really aren't going to get a good benchmark of compiler options that way. The real topic I guess, is how fast are the different distro's and why. We are off on a tangent that has probably been explored in detail by those GCC people :) I can still do a lot in assembler, which is basically knowing all the tricks of the architecture that you can find, so this facinates me in that respect, none the less.

Does anyone out there know if GCC has any reports on this kind of thing that they use themselves?

Re:The Mindcraft method, against itself

[Anonymous Coward]

O1 is the safe level of optimisation. Even O2 runs the risk of doing more harm than good, although it's a fairly low risk.

You're wrong. With gcc, -O2 is the safest since it is the best tested. By comparison -O1 is riskier.

Re:right now

[Rosco P. Coltrane]

Using it. It rocks. Best Linux distribution yet.

So are Debian, RedHat, SuSE and Slackware, according to Debian, RedHat, SuSE and Slackware users (respectively). I take it you're a Gentoo fan?

Makes an excellent point

[JanneM]

There are a lot of issues one can bring up with the test - not identical versions of various software; different X drivers, one distro will have patches missing in the others and so on. Clearly, that greatly influences the results.

And that is a good point to take home. Optimizing compiles is _not_ the panacea for speed and responsiveness that - a minority, I believe - of source-based distros tend to bring up. There are so many other factors intimately involved in it that any benefits are generally lost in the noise.

For some specific components, it can be a good idea - but for those, most distros tend to ship several optimized versions that the installer chooses between at installation time.

Another domain that benefits are specialized, compute-intensive applications; things like simulators or other technical stuff. But then, those apps are generally tweaked and compiled by their users no matter what distro is used anyway.

Re:Makes an excellent point

[Ruie]

You are quite right about the optimizations.

I would like to mention that the tiny cache size of the Celeron processor can very well make optimizations that increase code size have a negative effect.

/.ed text.

[Anonymous Coward]

The text, formating is a bit wack, sry.

Creating the Test

Obviously, the most direct way to test the compile time optimisations of Gentoo is to compile 3 Gentoo systems with different settings and then compare them. However, this really misses some of the story. Each distro has it's own attitude to the kernel and whilst it may be i386 or i586 compiled, it will have had some adjustments made to it. In reality, few people will be choosing to install Gentoo with less than recommended optimisations for their system. They are interested in a tradeoff between optimization and convenience. Thus, we aimed to compare Gentoo with an i386 based distro and an i586 based distro. With the assistance of some PLUG members we decided on Debian as the i386 candidate and Mandrake as the i586, in part as those were the options where people were available to do the install.

The following tests were outlined: Time to open a large sheet in Gnumeric. Time to perform a kernel compile. Time to perform "Duplicate Image" in Gimp. Time to perform a heavy "Unsharp Mask" in Gimp. Time to start OpenOffice "from scratch". Time to reload OpenOffice.

User experience to be assessed by all present on the day, using Galeon, Evolution, OpenOffice.

To make it easier to standardize for these tests we picked Gnome 2 as the Desktop Environment. This necessitated the use of the "Testing" flavour of Debian.

Hardware

The boxes from Evolution Xtreme had the following configuration:
Celeron 2 GHz Processor
256 MB DDR RAM
SAMSUNG - SP4002H 40G HD
MSI 6533E main board
All SIS chipset

lspci output:

00:00.0 Host bridge: Silicon Integrated Systems
[SiS] SiS651 Host (rev02)
[ 00:01.0 PCI bridge: Silicon Integrated Systems
[[SiS] SiS 530 Virtual PCI-to-PCI bridge (AGP)
[ 00:02.0 ISA bridge: Silicon Integrated Systems [SiS] 85C503/5513 (rev25)
[ 00:02.5 IDE interface: Silicon Integrated
[Systems [SiS] 5513 [IDE]
[ 00:02.7 Multimedia audio controller: Silicon Integrated Systems [SiS]
[ SiS7012 PCI Audio Accelerator (rev a0)
[ 00:03.0 USB Controller: Silicon Integrated
[Systems [SiS] SiS7001 USB Controller (rev 0f)
[ 00:03.1 USB Controller: Silicon Integrated
[Systems [SiS] SiS7001 USB Controller (rev 0f)
[ 00:03.3 USB Controller: Silicon Integrated Systems [SiS] SiS7002 USB 2.0
[ 00:0f.0 Ethernet controller: Realtek Semiconductor Co., Ltd.
[ RTL-8139/8139C/8139C+ (rev 10)

[ Installation
[ The 8139 NIC used the 8139too driver in all installations.

The Debian box was installed by Garry Buckle with aid from Matt Kemner. A standard Debian Testing setup was installed, but X was not persuaded to start with the stock SIS driver. As the stock kernel did not contain framebuffer support a new one (v2.4.21) was compiled to get video working. Upon testing with hdparm, it was apparent that this machine was having troubles setting above udma2. Eventually this problem was traced to the HD cable, a salutary lesson in the variability of identical hardware setups.

[ The Gentoo setup by Bill Kenworthy was compiled using the "stock" kernel source and the "-march=pentium3 -pipe -O3" compile flags. hdparm was needed to get dma on the ide running, despite it being in the kernel, but "xfree --configure" worked for Bill using the stock SIS driver. (Apparently the first time the command has worked for him!) The Gentoo install suffered a couple of false starts due to a typo using grub and OpenOffice was still being compiled the night before the test. 11 hours later the OpenOffice compile was still going and we thus had to regretfully abandon that portion of the test.

Garry's friend Joris (visiting from Belgium) put in the hard work on the Mandrake install, clicking "Yes" and "Next" like a pro to complete an impressively easy install. It defaulted to vesa framebuffer display, similar to the Debian install.

The Tests
[ Test 1 : Opening a 32,000 line sheet in Gnumeric.
[ We began this test with the mindset of testing

I use Mandrake.....

[MustafaJohnson]

But been thinking of switching to Gentoo to try it out. I was going to do that cause I thought specific compile might be faster than RPM......now I'm just confused. Anyone wanna put their two cents in on this?

Re:I use Mandrake.....

[Sloppy]

The reason for getting away from RPMs is when you see some software that you want, download the RPM, and you can't install it because it wants some other RPM that you don't have. Or worse, it refers to an RPM for a package that you do have, but it's a different version.

If you don't have that problem, then stick with what you have. Your life is good. Be happy.

(Back when I used Mandrake and Red Hat, I had that problem all the time and it was very frustrating.)

Re:I use Mandrake.....

[Ankh]

If you get dependency problems when running Mandrake Linux, you're probably not doing it right :-)

Seriously, the urpmi command will download an RPM for you, together with everything it needs that you don't already have, and install everything. Automatically.

See plf and the main urpmi web page for more info; Mandrake comes with urpmi, and it's also used by the GUI package manager (rpmdrake).

The trick to keeping Mandrake Linux running heppily is to avoid rpmfind. Always start by looking for packages in Mandrake's main and contrib (you can add them as sources for urpmi, and then it's automatic, just, urpmi package).

If you can't find a package, try searching with rpmdrake (you can search based on filenames too), or use urpmq -y some-substring such as, urpmq -y apache to see all packages that provide things containing that string.

if you use urpmi for package management, you should find that dependency and version problems becomre very rare. I won't say they'll go away (any more than they go away entirely with apt-get) but they will become very rare.

If you do get problems with urpmi, look in /var/cache/urpmi/rpms for the downloaded RPM files.

Gentoo loses!

[wirelessbuzzers]

... with -O2

Aren't most binary packages built with -O3?

Re:Gentoo loses!

[Anonymous Coward]

Absolutely not. Debian's packaging manual recommends -O2, but maintainers are free to choose some other optmization if they like.

You still have dependency hell

[grotgrot]

I tried Gentoo for a while and eventually gave up. The problem is that you still have dependency hell. Most packages look for stuff at compile time, and many have optional components. For example a video player may not include support for QuickTime unless the libraries are already on there at compile time.

So the fun starts when you start installing stuff, they don't include support for other components because they weren't there at compile time, you then discover the missing support, have to install the missing libraries and then recompile every package.

This is an especially big issue with multi-media stuff, and gets many layers deep as some libraries have optional components depending on other optional components.

About the only way to guarantee a fully uptodate system is to keep doing complete recompiles of the entire system until there are no changes.

Re:You still have dependency hell

[mark_lybarger]

yep, the portage system really needs some maturity that's obvious. i've had my machine unbootable after doing an emerge -u world. i think there was something about installing new glibc library and possibly removing the old one. i don't know, but i had to grab a knopix cd and chroot into the gentoo partition and basically re-install from scratch.

the portage also needs some sort of recursive uninstall. uninstall ide AND all the packages that depend on kde (or were built with kde support). uninstalling blackdown java should uninstall anything that required blackdown (and reset all the system variable crap).

that said, it's a great distro and i'll never find myself really using RH/ Mandrake or somesuch just because when i ran RH, i was always building kde from source, x from source and many many other packages to where it wasn't a RH system anymore.

and maybe the kewl gentoo users don't want a pretty gui installer w/ good hardware detection and all, but if one were put together, lots and lots of people would use it. gentoo is something that you install once and then just keep upgrading packages.

Re:You still have dependency hell

[GweeDo]

Looks like someone failed to set their USE flag properly. If you have it set right you will get support for all you want. Or if you do "emerge -vp packagname" before doing an actual emerge you can see what optional flags aren't getting used. People that use Gentoo but don't read the portage/emerge/use documents are asking for this. Gentoo isn't for all, it is only for the willing.

Please go here and reas as much as possible for installing Gentoo so you don't do something stupid.

Re:You still have dependency hell

[epyT-R]

That's what the --deep option is for ;)
you're missing quicktime support? np
just add it to your USE var and then rebuild the media player ebuild with -du. Sure, you still have to rebuild, but its automated at least.

A lot of distros ship applications with features missing, so I end up rebuilding half the system after install anyway(happened to me with slackware). With gentoo, I do it once, and that's it.

What portage does need is a USE var checking feature which would be run after a new value was added or removed (like nptl for instance). It would then scan the list of pkgs you have installed, and rebuild any that have nptl as a USE flag. This way, features can be easily 'turned on or off' with one simple change in make.conf.

Re:You still have dependency hell

[Anonymous Coward]

You totally missed what makes gentoo a great distro: the USE flags.

The USE flags are roughly equivalent to compile time options, like ./configure --enable-foo --disable-foobar

USE flags are set system wide in a configuration file (/etc/make.conf), and / or can be set before installing a single package
ex: USE="alsa -opengl" emerge xmame
will compile the package xmame without the opengl renderer but with support for the alsa sound system.

And portage takes your USE flags into account when it determines the dependencies for a package.
In the previous example, we requested alsa support for our program; if alsa wasn't already installed, then emerge will install it before compiling and installing the program that depends on it.

So what you're saying is totally wrong.
If you wanted multimedia applications that used the most common codecs, all you had to do was to add "quicktime avi mpeg oggvorbis" to your USE flags. The complete list of USE flags and their description can be found at: http://www.gentoo.org/dyn/use-index.xml. And there is no dependency hell, nor the need to recompile everything.

By the way, all of this is explained in the installation manual. A pity you skipped that part because it's what makes me like my gentoo distro so much.

Re:You still have dependency hell

[Anonymous Coward]

What whatever you wankers talking about "USE" flags and crap...

apt-get install <package>

Done... installs everything need for that package. Quickly too, none of that compiling garbage. And it works.

There is no Gentoo

Re:You still have dependency hell

[Anonymous Coward]

Watch this space for an excellent Gentoo installation tool. This guy is doing some pretty great work (pretty much by himself) making the gentoo installation quite literally a five step process. He plans on adding GUI support soon and gentoo has a hardware autodetection kernel used by glis, so you're set.

Re:You still have dependency hell

[SagSaw]

Hmm...I've never found this to be a problem in the year-and-a-half that I've used Gentoo.

I call bullshit. (Actually I don't think you read the instructions prior to trying to install Gentoo.) First of all, if qt has been in your USE flags since the start, everything you compile will have QuickTime support, if availaible. The problem might come in if you add qt later. Obviously, adding the qt USE flags doen't magically grant qt support to every package already compiled. The easiest way to find out which packages really need to be compiled is to read the ebuild for the program in question.

Of course, if you are using the ~ARCH option to automatically merge development packages, you probably will run into goofy packing dependecy problems at times.

One trick I use is to read the ebuild before merging a new version (not patch-level/revision) of a major packages such as X or qt. This lets me make sure I have the correct USE variables set for the features I want to use.

Re:You still have dependency hell

[HolyCoitus]

If you're talking about multimedia support, a lot of the problems with that have to do with the fact that you CAN'T package drivers for Quicktime and Realplayer directly. It's the exact same reason that Windows Media Player has dropped support. You get around this easily with the binary distros just because they package everything in there. That includes the Realplayer and Quicktime junk, but you can't download the mplayer or Xine source along with those codecs to my knowledge. I assume that is what you were meaning since you mentioned Quicktime in there. If there are other problems, I haven't ran into them yet with my Gentoo install. Any problem you may have anyhow is most likely documented someplace on the gentoo.org website anyhow.

Re:You still have dependency hell

[Delphiki]

You've actually got this backwards. In source based distribution they can not distribute source for certain things because of licensing. You can't compile GPL code with certain licensed code and then distribute it freely. What you can do though is distribute the code and let people compile it themselves, which is what Gentoo does. Gentoo tends to avoid a lot of these legal issues with the way that they distribute packages. Almost nothing is part of the actual distribution, it just exists on the mirrors and the repository contains instructions for the OS on how to build it.

Re:You still have dependency hell

[Anonymous Coward]

Whats the apt-get switch to install xchat without perl & python? Or kde WITH mp3 encoding support?

There is none. Gentoo gives you more freedom as regards to what you install and what optional features you want. With debian you get what the package manager choose. Nuff said

Re:You still have dependency hell

[HolyCoitus]

Yes, the way Gentoo comes it completely bare bones. It doesn't even come with X compiled after all the initial installs to get the main os running. With the media player though, it isn't so much dependency hell as it is that the open source mplayer can't distribute the propietary drivers along with it, so they have to be installed seperately to get the drivers that are actually going to work with those files.

Use flags could handle that possibly, but also to emerge realplayer you have to go download the realplayer package from their site because of the forms you have to fill out. There are a lot of annoying issues with the media players right now which I think make it difficult for a source distro to be able to distribute the proper codecs and drivers for the files to play properly.

Re:You still have dependency hell

[jimmy_dean]

I would take a look at the way you use Gentoo again. This is exactly what Portage is not. If you take a look at the USE flags for each package you want installed, you can see what it will be compiled with and what it will not. And if you do an emerge -p you will see the list of dependencies that it needs based on your current USE flags. Please learn how to use it properly and then you can critique it.

Re:You still have dependency hell

[Anonymous Coward]

Of course. Obviously something is wrong with your configuration. Which is never the fault of the system. Somehow.

You're right though, Gentoo isn't for everyone. Self-righteous pricks seem to like it though.

Oh yeah. Goatse for you, dickface.

*_g_o_a_t_s_e_x_*_g_o_a_t_s_e_x_*_g_o_a_t_s_e_x_*
g_______________________________________________g
o_/_____\_____________\____________/____\_______o
a|_______|_____________\__________|______|______a
t|_______`._____________|_________|_______:_____t
s`________|_____________|________\|_______|_____s
e_\_______|_/_______/__\\\___--___\\_______:____e
x__\______\/____--~~__________~--__|_\_____|____x
*___\______\_-~____________________~-_\____|____*
g____\______\_________.--------.______\|___|____g
o______\_____\______//_________(_(__>__\___|____o
a_______\___.__C____)_________(_(____>__|__/____a
t_______/\_|___C_____)/______\_(_____>__|_/_____t
s______/_/\|___C_____)__USE=_|__(___>___/__\____s
e_____|___(____C_____)\______/__//__/_/_____\___e
x_____|____\__|_____\\_________//_(__/_______|__x
*____|_\____\____)___`----___--'_____________|__*
g____|__\______________\_______/____________/_|_g
o___|______________/____|_____|__\____________|_o
a___|_____________|____/_______\__\___________|_a
t___|__________/_/____|_________|__\___________|t
s___|_________/_/______\__/\___/____|__________|s
e__|_________/_/________|____|_______|_________|e
x__|__________|_________|____|_______|_________|x
*_g_o_a_t_s_e_x_*_g_o_a_t_s_e_x_*_g_o_a_t_s_e_x_*


aslk;dnlnjkldfbjbhljkfh80t9-gqregn 0asasdf0 afjg89qarg ghasf8g9hd90sg fdsghlask;dfgnq88asdgzxcv zxlvkhqfi9ndbs 8asdg8asdgnkvnc dgjbi8bb787-asr f asdf afbn 8bf-f88a sfd8gh-a890sgsb4ta a tagsfbjizxpb8xczvnbrae ga08ergha08hgpxgng q0[azpsdklnvpalsdghgb08 arase8tghaklpcnvgnq3rqglskg8asdgnkvnc dgjfgnq88asdgzxcv zxlvkhqfi9ndbs 8asdg8asdgnkvnc dgjbi8bb787-asr f asdf afbn 8bf-f88a sfd8gh-a89bi8bb787-asr f asdg8asdgnkvnc dgjbi8bb787-asr f asdf afbn 8bf-f88a sfd8gh-a890sgsb4ta a tagsfbjizxpb8xczvnbrae ga08ergha08hgpxgng q0[azpsdklnvpalsddf afbn 8bf-f88a sfd8gh-a890sgsb4ta a tagsfbjizxpb8xczvnbrae ga08ergha08hgpxgng q0[azpsdklnvpalsdgdgn jasl;dkhn98as9b- zcgnasgmnzxcpvq3nnk as df;gkha;lnt4w,;n;gifgb8cx80v0z cased

Re:You still have dependency hell

[Anonymous Coward]

apt-get source xchat
# FIX IT YOURSELF HERE (find the configure line in the build script, or in most cases SET A VARIABLE, JUST LIKE GENTOO, OMG)
# build the package. debian/rules binary or something.

Repeat for the handful of packages that actually benefit from optimizations or absolutely need custom configuration.

And you're up and on IRC a week before you would be compiling everything from scratch. (Profit.)

Re:You still have dependency hell

[axxackall]

People that use Gentoo but don't read the portage/emerge/use documents are asking for this. Gentoo isn't for all, it is only for the willing.

I usually say: Gentoo is for users who can read.

Re:You still have dependency hell

[Anonymous Coward]

..and hurry on over to #debian on freenode to get insulted and humiliated by Raul and company... need that extra week to 'RTFM moron' and 'wtf do you want to do that'?

Debian.. the distro with an attitude!

Article missed the point

[Anonymous Coward]

The best way to optimize gentoo is too bootstrap it yourself. This yields impressive performance gains. In my case I settled on these flags:
-mcpu=pentium4 -march=pentium4 -mmmx -msse2 -Os -fomit-frame-pointer -pi
pe -fforce-addr -fforce-mem -ffast-math -mpush-args -mfpmath=sse

The other reason I use gentoo, is that it's easy to stay current. Once a week I do an emerge sync, followed by an emerge -u --deep world. The Gentoo package management system rocks.

Re:Article missed the point

[Anonymous Coward]

I think you'll find you forgot -O9 and -fomit-instructions. It might crash every now and then, but I swear bash is responsive in ways I never thought was possible.

Re:Article missed the point

[lanalyst]

O9 with gcc 3.x?

Matches my real world dual-boot experience

[Captain Kirk]

I dual-booted Debian and Gentoo thinking I would migrate completely to Gentoo for desktop use and Debian for servers. Galeon on Debian was way faster. In the end, I got fed up of compiling and re-compiling X and stuff trying various gcc switches. Debian is fast enough to make sitting about wiating for stuff to comlile a waste of time. And apt-get is every bit as good as emerge.

great article!

[randyest]

But I think I've read it a few times before already . . .

Here's a complete repost of the text in case it gets slashdotted. Hmm, guess I should post as AC, but wtf. Just don't mod this up.

Warning: Lost connection to MySQL server during query in /home/misskim/public_html/linmagau.org/pnadodb/dri vers/adodb-mysql.inc.php on line 170

Warning: MySQL Connection Failed: Lost connection to MySQL server during query in /home/misskim/public_html/linmagau.org/pnadodb/dri vers/adodb-mysql.inc.php on line 170>
mysql://misskim:@localhost/misskim failed to connectLost connection to MySQL server during query

Hmm, guess I should post as AC, but wtf. Just don't mod this up. :)

Interesting, but I have to wonder.

[Enahs]

For me, Gentoo is a great choice partially because I like the control and partly because I use crufty hardware that doesn't fall into any predefined (read: Intel) category.

Try using binaries compiled for an i686 on a Via C3-1G, for example.

Yes, if your entire reason for using Gentoo is to have control over how apps are built, starting from stage3 pretty much defeats the purpose, and yes, if you don't know what you're doing, then rebuilding X can be a real drag. However, I have to say that I appreciate the fact that Gentoo manages to avoid a lot of legal issues by having the user build the packages her/himself. Honestly, I'd love to be Ogg Vorbis-only for music on my computer, but when I own a portable MP3 player, an MP3-capable DVD player, an in-dash MP3 player, and use OS X at work where QuickTime Ogg Vorbis support is dodgy at best, I want lame. And I want lame support built into kdelibs or whatever lame support needs to be built into so that I can drag-and-drop 192kbps ABR MP3s from an audiocd:// ioslave window to my mp3 folder. ;-D

My own experience has been that Gentoo outperforms Debian on my hardware, but only after I've done some tweaking on Gentoo. YMMV.

Bleeding Edge

[Ikkyu]

I got hooked on gentoo because of the bleeding edge factor, I usually have apps built and installed on my machine within a day or two of revs. That is just somehting that I never found with Mandrake.

Re:Bleeding Edge

[buchanmilne]

That is just somehting that I never found with Mandrake.

Then I take it you were running a stable release? Popular software in Mandrake is updated very quickly in cooker (unless there is a good reason not to, such as OpenOffice.org is a point-point release behind because the maintainers working on 1.1).

Re:Bleeding Edge

[Ikkyu]

It's not that I couldn't find the package I wanted, it's that I was relegated to special level of hell where you must resolve dependencies for binary packages.

Re:Bleeding Edge

[buchanmilne]

It's not that I couldn't find the package I wanted, it's that I was relegated to special level of hell where you must resolve dependencies for binary packages.

???

urpmi? Or were you trying to use rpm itself?

It could have been worse, you could have been trying to resolve source dependencies, having to rebuild a whole chain of software with new options to get one more feature that's just one package away on a binary dist like Debian or Mandrake.

Re:Bleeding Edge

[HBI]

You can get _too_ bleeding edge with Gentoo.

My Gentoo partition is out of action right now because Portage failed to update the glibc dependency while doing other updates. Yeah, I know, I can fix it manually but it's not worth my time at the moment.

Put a real damper on my desire to run Gentoo anymore to tell the truth. Up to that point I was a real believer but I got fucked over just the same way that the RPM distros fuck you over, by ignoring dependencies or putting you through hell getting them up to date.

Maybe I should just bite the bullet and run BSD.

Re:Bleeding Edge

[Sebby]

I've had a similar thing happen on a RH Linux update using their 'up2date' thing - but this was for a kernel update, on a production machine.

NOT fun.

Since the system was remote, we couldn't work on it directly; down for 2 days and a nice figure in support costs.

When we're due for a server change, I'll recommend an Xserve. At least those don't need constant updates (of course, they might not be 'bleeding-edge' secure either because of that), and I've never had/seen/heard of an update problem with them.

Re:Bleeding Edge

[ldspartan]

Lord almighty... you had a machine in a production environment, and you did a kernel upgrade with an automated package management tool?

I hope no one is paying you to administer things. I can perhaps understand doing a remote kernel upgrade on a machine you don't have ready physical access to if you're compiling it yourself, but it certainly didn't take me long to figure that _no_ package management system should be trusted with making changes to the kernel.

I guess you have to learn things somehow, but christ...

Re:Bleeding Edge

[Sebby]

I sure did: do not trust RedHat :)

Mind you, it didn't totally hose the system (could still boot from the old kernel, if you had physical access), but their tool was definitively broken, and the hosting company's support personel probably didn't help matters....

I'll tell you one thing: I had a few choice words for RH ;)

Re:Remember the KDE mandrake/gentoo fiasco?

[Anonymous Coward]

This (I think) Is because Mandrake runs it's X server with a priority of -10 (high priority). Gentoo, by default, runs X at priority 0 (normal priority).

One more thing

[Enahs]

From the article:

Upon testing with hdparm, it was apparent that this machine was having troubles setting above udma2. Eventually this problem was traced to the HD cable, a salutary lesson in the variability of identical hardware setups.

Very telling pair of sentences.

Unfair test

[periscope]

There seems to be little attention given to the fundamental unfairness of this test presented.

The distributions were running with different software versions initially and although this was corrected there seems to have been little consideration given to the minor tweaks given to each different installation used. Which services were running on each system? Were the kernel settings identical in use? Were the machines experiencing differences in performance due to the X setup causing X to add different loads?

etc.

Fundamentally this test was probably not complete enough to suggest anything in particular. Perhaps it would have been better to boot a single machine three times and perform the sequence of events exactly the same each time as this would have also ruled out some other potential factors.

Jon.

Re:Unfair test

[Anonymous Coward]

Fucking hell... we've had to put up with wannabe 1337 assholes banging on about compiling from source with Gentoo for months. We've heard all kinds of "it's faster, it's easier" from stupid propaganda spewing zealots in every article from lego to election fraud.

1. We know it ain't faster. The test proves it and rather takes the wind out of the Gentoo zealot sails.

2. We already know it isn't easier.

So Gentoo users... fuck off, ok. Your little day in the sun is over.

Re:Unfair test

[Boglin]

An owner of a Hummer once told the owner of a motorcycle that his Hummer got better gas milage. The motorcycle owner said this is ridiculous, and challenged him to a contest. The would each drive the same hundred miles of road and see who used more gas. They did, and the motorcycle used less gas. The Hummer owner immediately pointed out that the test was unfair "Your motorcycle weighs far less than my Hummer. That's part of the reason you're getting better milage. Also, you're using a different engine with less horsepower. If your motorcycle weighed as much as my Hummer and used the exact same engine, you would see that I get better milage"

You see, part of the reason that you would use Debian or Mandrake is that the developers have gone in and done those little tweaks. You are trusting those developers to have compiled a faster kernel and to turn on only the correct number of services. With Gentoo, you are claiming that you can do a better job yourself.

I also wouldn't be overly surprised if they were using different versions of the X libraries. After all, several of the other poster here have talked about how they loved that Gentoo always gave them the latest and greatest libraries, many stating that they do weekly emerge syncs. The binary distro users, on the other hand, don't have that option and usually use the older binaries for a while. Therefore, to have the Gentoo machines and the binary ones run the same versions of the libraries assumes that the Gentoo user hasn't been updating his system to the new libraries, which doesn't seem to be a realistic expectation in practice.

As for whether the machines were running the same services, binary distros tend to run a lot of them by default, Mandrake in particular. Gentoo, if I understand correctly, usually wants you to choose which services to run. I highly doubt that the testers would have turned on more services than necessary, so, if anything, Gentoo has the advantage on this front.

I do agree with you that they should have ran each of the tests several times to make sure that there were no flukes. But to say that the test is invalid because there were differences between the systems is to miss the entire point of the test.

Re:Unfair test

[XO]

That was the WHOLE POINT in case you missed it: not to test identical machines with virtually identical software.

The idea was to test flat out time it takes to do something, on mostly stock versions of the varying distributions!

It's meant to test how unfair the difference would be between distributions!!!! Therefore, it is fair.

Get a grip, people. :-)

Re:Unfair test

[sankeld]

Fundamentally this test was probably not complete enough to suggest anything in particular.

It's intention seemed not to compare program versions or compilations, but vanilla installations of various distributions. I am led to believe that most Gentoo users set up their system the same way this article outlined. And most, would likely assume it is faster than Debian or Mandrake installations.

This article clearly shows how this is not the case.

Oopsie!

[Enahs]

Apparently DMA is enabled on all the machines. Missed that on the first read. Also see my other comment on the article: apparently they ran into differences in the machines, despite the fact that they were supposed to be identical.

Despite missing an obvious point, I still stand by my original sentiment: this article isn't very informative at all.

Re:Remember the KDE mandrake/gentoo fiasco?

[Sir Joltalot]

Myself, Gentoo's biggest feature was the kernal compile options, adding patches for pre-emptive mulitasking, and improved responsiveness.

Really? Hmm.. I run Gentoo and I use vanilla kernels because I find they perform better. I'm on an SMP box though, so that might have something to do with it. But I tried a couple Gentoo kernels and I had *seriously* bad performance problems. Whenever I was compiling the mouse cursor would get all jittery, as would the scrolling song title in XMMS - even if I niced the emerge down to 12 or something!

Vanilla kernels tend to run really nicely for me and if I nice the compiles down I barely even notice them (even if I don't, I barely notice them, being on an SMP box :)

Maybe I'm choosing bad options in the Gentoo kernel though; do some of the patches (i.e. low-latency and pre-empt) interact badly or something? What options do you use for your Gentoo kernels? Just curious.

One thing the Gentoo kernels are good for, though, is Starcraft in Wine. Dunno what it is about them, but man Starcraft sure runs really fast with those kernels. Vanilla kernels run it really, really slowly. At least in recent versions of Wine; 20020510 was the sweet spot for Starcraft, IMO, but it's incompatible with glibc 2.3 and thus no longer an option. Sigh.

compiler comparison

[pimpinmonk]

i'm glad your super-special system with an entirely different version of gcc can compile faster than my vanilla gcc system, but that's not how this test was conducted.

Hardware-wise it was the same system for each distro, and they listed that mandrake and gentoo are running the same exact version of gcc. So Obviously although gentoo is great, Mandrake's got some optimisation tricks thrown in to speed certain things, be them library or kernel-wise.

The point was also to compare relatively stock installs. Sure, you can argue that you can further optimize gentoo, but you can do the same for Mandrake and Debian. This test compared the out-of-the-box performance of Debian, Mandrake, and Gentoo to see if the optimisations that most Gentoo users are running (which just happen to occur "out-of-the-box") really impact the system speed.

Try it againts Red-Hat, the "Real Man's" OS

[Anonymous Coward]

(eom)

Reasons I like Gentoo

[caluml]

I like:
1. The very large collection of packages. For example, I don't believe you can apt-get install vmware.
2. The very sensible defaults. After emerge wine, I could run Lotus Notes 5 with no changes at all. I tried to set that up on a RH box with wine compiled from source, and it chucked up loads of errors and didn't work.

That said, it does take a long time to set up from stage 1. You're probably looking at about 3-4 days for the base system, X, KDE, Mozilla, and OpenOffice. I'll use it for my personal machines, but I can't be arsed with all that for run of the mill boxes.

Happy as a wet turtle Gentoo user

[GweeDo]

I have been using Gentoo for months now and will never turn back. Little of this has to do with performance and 99.9% of has to do with Portage. Package management, dependency checking and the lot are SO great. Secondly is where performance comes in. Without proper CFLAGS you might as well ignore this. On my Athlon XP 2800 I have this:

CFLAGS="-march=athlon-xp -O3 -pipe -fomit-frame-pointer -mmmx -msse -m3dnow"

In some simple tests I have done I have seen this as worth while. I have two pages I have created that might be worth a read:

CFLAGS Guide
Is -mmmx and such worth it?

Hope you enjoy these reads.

Re:Happy as a wet turtle Gentoo user

[Anonymous Coward]

I want compilers to compile just int main(int c, char **a) { return 0; } when they encounter benchmark programs such as yours...

Re:Happy as a wet turtle Gentoo user

[EvilTwinSkippy]

I second that. I'm running Gentoo on all of my datacenter servers. I'm not as concerned about performance as I am the ability to preserve the operating envirnoment of the machines between OS upgrades.

It takes months to get a mail server properly tweaked, or the delicate Apache installation operational. It really sucks to sweat bullets between living with a root-exploit, trying to re-synthesize RedHat's configuration from source, or praying that everything still works after doing an OS upgrade in place.

These machines need to operate for years at a time, and as frontline webservers, email, and DNS hosts you can't afford to be behind on your patches.

Re:Happy as a wet turtle Gentoo user

[antiMStroll]

Why do you disable SSE, MMX and 3DNow support? That seems counter-intuitive.

Re:Happy as a wet turtle Gentoo user

[Anonymous Coward]

Uh, you might want to look into Debian.

No time wasted compiling and the dependency system is way better than Gentoo. Plus stuff "just works", unlike Gentoo which sometimes won't compile.

There is no Gentoo

Re:Happy as a wet turtle Gentoo user

[GweeDo]

-mmmx -msse -m3dnow That makes sure they are enabled :) - doesn't = turn off. It is actually a bit redundent since my -march flag turns them on anways :)

Re:Happy as a wet turtle Gentoo user

[Anonymous Coward]

> Package management, dependency checking and the lot are SO great.

Dependency check and package management? WHAT A NOVEL AND UNIQUE IDEA!

Re:Happy as a wet turtle Gentoo user

[lewp]

A datacenter full of Gentoo? Sounds remarkably like the nightmare I've been having for the past few months. You poor soul, at least I wake up eventually.

Re:Happy as a wet turtle Gentoo user

[Fnord]

Your "Is -mmmmx and such worth it?" guide is a little unfair. Thing is, you notice that only -O3 really made much of a difference. Well, that's because each of your tests is just one big loop and -O3 does heavy loop unrolling. You basically chose the absolute optimal case for -O3 to win (besides possibly having a small function call inside that loop so that -O3 could inline it). And you didn't do any floating point multiplies or divides which is where -mmmx+sse and -m3dnow would help you.

If anything you were also using a relatively small dataset. If you get a large enough data set (or code size) -O3 might actually hurt you (loop unrolling and function inlining will bloat both code and data size and make it much more likely to have a cache miss).

Anyways, synthetic benchmarks are one thing but your is so synthetic as to be rediculous.

Re:Happy as a wet turtle Gentoo user

[paxil]

GweeDo wrote:


In some simple tests I have done I have seen this as worth while. I have two pages I have created that might be worth a read:


I took a look at your pages. Particularly interesting is the page with the benchmarks. You had better take a look at the code gcc is generating with these options. The reason -O3 gave a factor of 20 speedup with your simple tests is because gcc is smart enough to tell that your code never does anything with the variable named "something." Because this variable is never used, gcc eliminates all of the arithmetic involving this variable when you ask it to optimize. The inner loop is completely eliminated from your code when you pass gcc -O3.

Re:Happy as a wet turtle Gentoo user

[Anonymous Coward]

Actually -mmmx, -msse, and -m3dnow simply turn on the __builtin_ functions that interface the relevant instructions. This is for developers, and if they do not use them, they will not do anything. Now -fpmath=sse will change the code generated by GCC, but -msse will not. In Gentoo those flags will be used by packages that actually would make use of them, presuming of course that you have the relevant USE parameters. Putting them in your CFLAGS is simply stupid.

Re:Happy as a wet turtle Gentoo user

[Chreo]

On my Athlon XP 2800 I have this:

CFLAGS="-march=athlon-xp -O3 -pipe -fomit-frame-pointer -mmmx -msse -m3dnow"

Um, did you know that -march=athlon-xp implies that gcc should include MMX, 3DNow and SSE optimisations by default and thereby rendering the switches -mmmx -msse -m3dnow redundant?

This goes for all -march switches so by using -march=p3 you get -mmmx -msse for free.

Oh dear, well now you know!

Re:Happy as a wet turtle Gentoo user

[The Analog Kid]

I've noticed that -mfp-math=sse,387 gives even faster performance.

Re:Happy as a wet turtle Gentoo user

[Arker]

If you're compiling your whole system with those flags you are really screwing the pooch when it comes to performance.

O3 gives a great performance boost for certain bits of code, such as your benchmark. It's very counterproductive for other things, however. It turns on things like inlining functions that are not turned on at O2 for good reason. In many cases they do more harm than good.

O3 should only be used when you're dealing with a relatively small program and you can test to make sure it's actually helping. Do your whole system with it and it's going to cause a drastic increase in cache misses, more than outweighing the better performance in certain code segments.

Re:Happy as a wet turtle Gentoo user

[Anonymous Coward]

I second that. I'm running Gentoo..

Why is it that all the people who don't understand Gcc flags are Gentoo users?

..on all of my datacenter servers.

In the interests of discloser, please tell us all what company you work for so we can avoid it like the fucking plague. I want to know that the person looking after my purchase information is actually competent and understands what they're doing. In other words, not you!

If you think that is harsh, it is. Go read a Gcc manual and maybe you'll understand.

Re:Happy as a wet turtle Gentoo user

[EvilTwinSkippy]

I'm sorry to hear you've had a bad experience.

For what it's worth, I've managed to make it work, and work quite well.

But then again, I compile all of my machines with a vanilla set of optimization flags. I just use Gentoo for the portage package management system. I also have one machine I do all the testing and upgrades on before I ship that portage tree to the rest of the servers.

Gentoo isn't particularly good or bad. It simply requires a higher level of intimacy with your system. Besides, you DO know everything that is installed on your servers? Don't you?

Diminishing returns

[zsazsa]

From the article: The Gentoo install suffered a couple of false starts due to a typo using grub and OpenOffice was still being compiled the night before the test. 11 hours later the OpenOffice compile was still going and we thus had to regretfully abandon that portion of the test.

So when does the time taken to compile the app with extra optimizations exceed the time you save on tasks performed in that app? Of course, that's only if an optimized build is faster, which in this tests did not appear to be the case, compared to Debian's oldschool 386 builds and Mandrake's 586 builds (by the way, does Mandrake still give you 586-optimized binaries by default?).

I also wonder why they used -march=pentium3 instead of -march=pentium4. Do they not know that the new Celerons are P4-based? Is -march=pentium4 too buggy? This article leaves a lot unexplained and doesn't seem very scientific.

Re:Diminishing returns

[Rhone]

So when does the time taken to compile the app with extra optimizations exceed the time you save on tasks performed in that app?

Does it matter when you're letting stuff compile overnight, or while you're at work/school/whatever? Or in the background during times when you're doing something that isn't CPU-intensive?

I mean, it's not like Gentoo users sit in front of their computers twiddling their thumbs, just waiting for a compile to finish before they can move on with their lives.

Speaking of arrogance...

[Anonymous Coward]

Have a nice day. And try to learn, this time.

I can help you with your punctuation if you'd like...

Mod this up

[Anonymous Coward]

So that the crybaby "I'm a real Linux user because I don't use Redhat or Mandrake" Linux users will realize that Linux is Linux, no matter what distrubution you use.

I tried Gentoo. It wouldn't compile on my k6-3. Put in on my Athlon 700, expecting substantial increases in speed. Boy was I wrong. KDE still took the same amount of time to load. Mozilla still took 5 seconds to load. Gimp was not any faster.

Re:Mod this up

[archen]

"I'm a real Linux user because I don't use Redhat or Mandrake"

I think that additude is more because those distro's tend to hide a lot of the fundamental parts of really administering a Linux box. Mainly because there seems to be a growing ammount of people who can use Linux, and perhaps even get around okay, but have never even compiled/installed from source before.

In some ways it is an elitist additude, and in some ways it's sort of a valid concern. It's also a problem because there's a lot of nice software out there with no binary form (rssh as one example). Not that it's hard to check an MD5, and read a readme file - but some have never done it and you wonder what other areas they're probably lacking in understanding as well. Like a windows admin who's never used regedit - using it may not be neccesary on a regular basis, but not knowing how to use it is probably shifty at best.

Re:Mod this up

[Anonymous Coward]

http://packages.debian.org/unstable/net/rssh.html

Re:Mod this up

[Anonymous Coward]

I think that additude is more because those distro's tend to hide a lot of the fundamental parts of really administering a Linux box.

You're right, and this is an elitest attitude which just proves that anyone who tells you Mandrake or Redhat arn't "real" Linux distributions are generally idiots. So I use Mandrake because I don't want to waste days at a time configuring everything. I can change my video setup in Mandrake using the Control Panel instead of fidling with Xf86Config. Horror!

Yet just yesterday I configured X on my second box to export the display to my main box. Using xauth & writing my own .telnetrc

Its not that I can't configure a Linux box if I need to. Its just that I have better things to do. I guess configuring Linux must be a hobby for some people though.

Re:Remember the KDE mandrake/gentoo fiasco?

[mickwd]

It's not part of the main distro, but there is a kernel-multimedia-2.4.21.0.16mdk-1-1mdk.i586.rpm in Mandrake contribs. Check it out if you want a more responsive kernel.

dosemu and WP 5.1 ?

[Anonymous Coward]

Does dosemu let you run WP 5.1 so that all the F keys work ? Or is that another configuration you have to track through ancient usenet posts ?

Ob. Translate-o-matic post

[tempest303]

Here you go, the obligatory Gentoo Zealot Translate-o-matic reference!

Enjoy!

Re:Ob. Translate-o-matic post

[Anonymous Coward]

OTOH, the Redhat-translate-o-matic is very simple:

1 I'm running Redhat. Gee, I must be a professional, so I'm right.

2 A installation problem with a RH system is always the fault of a third party. A driver, non RH package etc. It is never the problem that the Redhat delivers too little packages, and too old, so you MUST use that third party package to get things running. (SUSE does this much better). Anyway, in case it might be such a problem, rule #1 applies.

Re:Ob. Translate-o-matic post

[tempest303]

SuSE is also non-Free, so what do you expect? They're allowed to get away with things Red Hat can't, being a 100% Free Software distro.

Also, the "too old/too few packages" problem is now on its way to being solved.

Finally, it doesn't take a "professional" to desire a sane distro. ;-)

Re:Ob. Translate-o-matic post

[Anonymous Coward]

Note: I am not the author of this fine little rant - I found it on Slashdot, but it was posted anonymously. I edited it a bit, and have put it up here for posterity and the enlightenment of all.

Hey, you stole my fine little rant! You bastard!

Mirror

[Anonymous Coward]

The article seems a little slow from where I sit. Here is a mirror.

Look, Maw, I'm feedin' them trolls!

[Enahs]

That's because you lack experience to read between the lines.

I did, and seriously doubt this could be a fair test. The only way you could completely miss the point I made and then make it again for me is if you're either trolling or forgetting to take your medication. Which is it?

Have patience and in some 10 years you will be able to make much of even a short, beautifully written article like that.

In 10 years I've seen tech writing go from "let's praise anything Microsoft" to "let's bash anything that's popular." The writing hasn't improved; it's gotten worse. Even more dismal is that testing methods get worse and worse.

Read the article. The reviewers even acknowledge that their identical systems aren't identical!

You'll also become less arrogant, which may help you deal with people.

I'm so glad to see that an Anonymous Coward knows so much about me.

Have a nice day. And try to learn, this time.

Thanks, for your gratuitous, comma.

Re:Look, Maw, I'm feedin' them trolls!

[arkanes]

The "difference" was a hardware fault in a cable. They mentioned it as a reminder that it's important to be aware that even nominally identical hardware can have hidden differences (like failures). I'm sure the cable they replaced (did they eventually replace it and get DMA working?) would have been the same cable the others used.

Photo views a little skewed?

[SassyDave]

Does it seem strange to anyone else that in the linked photo gallery, the only picture with a female has been viewed 3 times more than any of the other pictures?

Sheesh.

I guess there's just not much scenery to show off at distro day.

Re:Photo views a little skewed?

[IIRCAFAIKIANAL]

C'mon dude, she's hat!
(no, I spelled that right. Say it out loud. Add a bit of Cartman to your voice as necessary.)

Besides, i'm sure now it's going to hit 20x as much thanks to you :)

...and 11 hours later:

[Anonymous Coward]

OpenOffice.org!

Re:...and 11 hours later:

[SimGuy]

There are advantages to emerging the openoffice-bin package. :)

Default SiS driver sucks

[uncleduck]

I switched to Gentoo because it was the only distro (tried several versions of RedHat, Mandrake and even the latest Slackware, have'nt tried Debian yet) that would give me any kind of performance on my useless motherboard with an onboard Sis 530 Graphics Chip. Even Gentoo was kind of slow until I used the Sis drivers from http://www.winischhofer.net/linuxsisvga.shtml. This guy has a ton of really useful info for people with any of these old junky graphics chips and no budget for real equipment. The difference for me was night and day. I tried the new drivers in RedHat and Mandrake. Although there was an improvement, it was nothing like the improvement for Gentoo! The lesson of all this? I will never buy onboard video again! My next box gets Nvidia or something else that works!

The real advantage of a source based distro

[Distinguished Hero]

might be seen when AMD finally launches the Athlon64. Compiling everything to x86-64, and thereby increasing the numbers of registers substantially, as well as migrating to a slightly cleaner ISA (sorry x86-32 lovers) should render better results than attempting to optimize for a Celeron (ewww!).

On a less serious note, the server seems to be having a little bit for trouble; maybe they attempted to install Gentoo on their server (j/k).

Re:The real advantage of a source based distro

[fok]

This is true, and I agree. I like to compile everything to optimize performance. BUT, when I buy my new Opetron, my kernel will probably not run, and most of the applicantions I've compiled will be rendered useless (I think...), so I'll have to start it all over. This will (I think...) not hapend if I compile it for 386 for forward compatibility.

No. Gentoo is not a performance leader.

[0x0d0a]

Having said that, it looks like the guys doing the testing got their CFLAGS wrong. Gentoo's performance should never be worse than Mandrake -- I reckon they forgot omit-frame-pointer.

Omit-frame-pointer is not a regular optimization. Working without stack traces to hand to a developer if you have a problem isn't really a reasonable optimization unless you're doing something like an embedded system, where you couldn't get at the stack trace anyway.

This is *exactly* what the real tech-heads have been saying for years, what my tests confirm, etc. A minor change in a couple of compile flags above -O2 almost *always* makes very little difference. Compiling your own packages really just plain doesn't matter. Maybe if gcc really was incredibly tuned to each processor, but certainly not with the current compiler set.

Also, the kernel compile is unfair, because gentoo-sources includes a whole load of patches that Mandrake and Debian don't.

And perhaps the inverse is true, too?

Look, the point is, Gentoo is not significantly faster than any other general distro out there. If you use it, it's because you like their tools or packaging scheme. You aren't cleverly squeezing out more performance.

Oh, and last of all, I've seen compiler folks saying that it's not that unusual for -O3 to perform worse than -O2. When I was taking our cache performance analysis bit in university, cache hits and misses really *is* the dominant factor in almost all cases. Loop unrolling and function inlining can be a serious loss.

Finally, compiling for different architectures generally makes very little difference on any platform other than compiling for i586 on a Pentium. The Pentium runs 386 code rather slowly. The PII and above will happily deal with 386 code.

I don't mind compiling times

[The Analog Kid]

Since I do my compiling while at work and during when I sleep. I also use distcc. I also like how there is no dependency hell, and I know whats in my Gentoo system, with no programs I don't use.

Catch 22

[Alethes]

It seems that the people that would benefit the most from a source-based distro and optimizing binaries specifically for their hardware are the ones with the slow hardware that will take too much time to get everything installed for it to be a worthwhile investment of time.

Re:Catch 22

[stang7423]

For those of you installing gentoo on slow hardware here are your installation instructions:

1. %emerge system
wait 24 hours...
2. %emerge "all packages you want"
wait 24 hours...
3. Profit ??? your system is complete.

For those of you that say compiling on slow hardware isn't a worthwhile investment of your time, stop watching every line of code compile. Your computer is a big boy and he/she can operate for hours on end without you looking over its shoulder.

Re:Catch 22

[Anonymous Coward]

Heard of distcc and ccache? there's also emerge --buildpkg and quickpkg for creating binary tarballs on another system.

Confusing how a wild generalization that a group of users runs slow hardware rates an 'Insightful' - I suppose you verified your info at stats.gentoo.org?

gentoo forums responses

[zojas]

What people are saying on the gentoo forums about this too-brief article.

Re:Remember the KDE mandrake/gentoo fiasco?

[BrookHarty]

It's not part of the main distro, but there is a kernel-multimedia-2.4.21.0.16mdk-1-1mdk.i586.rpm in Mandrake contribs. Check it out if you want a more responsive kernel.

Wow, thank you, didnt know about that kernel, looks like it has the patches I was talking about. Did a quick lookup on pbone and found the info on it.

This kernel includes patches useful for multmedia purposes like: preemption, low-latency and the ability for processes to transfer their capabilities. The preemtion patches allow a task to be preempted anywhere within the kernel, using spinlocks as markers for non-preemptibility regions. The resulting system response is greatly increased, with measured average latencies under 1ms. Andrew Morton's low-latency patches fix the remaining points in the kernel that cause latency. The setpcap patch allows suid root processess to transfer capabilities to non-root processess, and so making it possible for user processes to run with realtime priority.

Gentoo vs. RedHat 9

[SassyDave]

In my experience, Gentoo "feels" a whole lot faster on my laptop (P3 600Mhz, 160Mb) than RedHat 9 does. Gentoo, however, will clobber your system while doing updates. I much prefer apt-get update over emerge sync, simply because the rsync that powers Portage takes a very long time to update your portage tree on semi-old hardware. People have told me to "just run the updates at night", but doesn't that take all the fun out of it? For having the most up to date packages, all the time, Gentoo is great. It comes at a price, though. One thing is for sure: Your machine will never work as hard with any other distribution but Gentoo.

Gentoo is truly amazing

[Anonymous Coward]

I can't recommend Gentoo enough. After spending under seventeen weeks compiling the system under several combinations of flags such as "-mathlon-xp", "-mprocessor-serial-number=948FCDE", "-O42", and "-fomit-stack-pointer" I have finally found the perfect combination that gives me a 1.5% performance increase while running my two most commonly used applications: the "joe" text editor and, of course, BitchX. And now thanks to the magic of Portage, I can reproduce these dramatic performance improvements on any system, as long as it is identical in every way to my own and I'm willing to sacrifice four-to-five days to a full compilation from source.

Wait to go, Gentoo. You can't ignores the savings provided by target-platform specific compilation!

Re:Gentoo is truly amazing

[Anonymous Coward]

I think you forgot something. Since you're running gentoo, don't you need to have an optimized build of wget running in the background reloading distrowatch repeatedly, so that Gentoo can stay the most popular distribution?

Re:Remember the KDE mandrake/gentoo fiasco?

[Enahs]

0 is where it should be. Other distributions run X at a higher priority to make up for "vanilla" 2.4's crappy interactive performance. Once 2.6 is out, this won't be an issue anymore. Also Gentoo's gaming-sources and ck-sources (my personal favorite) have optimizations built in that improve interactive performance greatly, eliminating the reason for running X at a higher priority. Some people report having problems with X running at higher priority; I never have (some people have problems with soundcard starvation, among other things) but then again I don't have to worry.

To be fair, I'm running RH9 until sometime tonight (have a chrooted Gentoo build waiting to be installed) but I'm running a Planet CCRMA kernel, which includes a number of the ck-sources patches.

Debian alien

[KMSelf]

VMWare on Debian is best accomplished with alien to convert VMWare's RPM to a .deb, and installing this via dpkg.

As for sane defaults -- Debian tends strongly toward this in my experience.

Re:Debian alien

[caluml]

VMWare on Debian is best accomplished with alien to convert VMWare's RPM to a .deb, and installing this via dpkg.
vs
emerge vmware

No comparison ;)

Re:Debian alien

[Anonymous Coward]

When I used debian, I wanted to keep my system clean and when I installed vmware I really anted to convert it to a deb with vmware. However, for some reason, the installed vmware from the deb wouldn't work at all, and I resorted to installing the tar.gz by hand, which worked fine.

On gentoo, emerge vmware-workstation keeps my system clean, and, it works

Re:Debian alien

[Anonymous Coward]

Ahh but will VRMS rag on you to the ends of the earth if you do?

alien? Love it!

Gotta love that freedom...

Why I like Gentoo

[SpineZ]

I've tinkered with various distributions over the years and the main reason I like Gentoo is the package management system, Portage. It is by far the best package mgmt system I've ever used in a Linux environment. The thought of RPMs frighten me.

It's hard to say if it feels more optimized than say RH 9 or the latest Mandrake, but it's perfect for my needs as a desktop system. I am able to play Warcraft III under Transgaming's WineX without a hitch.

Also, with the install process and actually compiling from sources, you will learn more about linux and how it works in the first week than in the first few months with other distros - my opinion/personal experience of course.

Sites to help you get started:
Gentoo Installation Guide for x86
Optmized CFLAGS for compiler - Be wary of Pentium 4 options. See forums for more information.
Gentoo Forums - arguably the best linux support site on the 'net

Re:Why I like Gentoo

[Gta-Klue]

While the speed of programs compiled in portage is an important factor, my desktop seems much more responsive than it was under RH or Mandrake (two other distros I tested before my recent switch to Linux), it is not the selling point for me.

and It is the ease of installing programs that portage brings

Ok, on the first.. I've found a direct correlation with the responsiveness of the desktop to which WM your running.
Even my eMachine 400hmz celeron was slow with Mandrake 9.0 and Redhat 8.0 running KDE or Gnome. But load it up with Blackbox/Fluxbox/XFCE etc.. and the desktop is as snappy as ever. (Not to flame, just an observation)

On the second.. apt-get install is just as easy. :)

I'm really suprised no one has mentioned this yet, but with the amount of computing power you can buy right now with very little money, is it the distro that's faster, or is it the hardware? I know at some times my Win* boot is a little faster copying files compared to my Debian boot, but then the hard drive is a faster RPM one than my debian.

Having NOT read the article, but having had experience installing and compiling Gentoo, I just do not have the patience to spend a day or more downloading, compiling and installing just the absolute basics, when I can insert, boot and install Knoppix.

With my 1.8ghz P4, 512mb ram and 64mb Nvidia GForce card, Knoppix/Debian runs just fine for me. But, I can see why someone wanting to salvage an older P2, or P3 system would want to do a Gentoo install. Especially if it's not thier main boxen. Since mine is my ONLY boxen, that much down time (and lack of /. would drive me nuts ;))

Re:Why I like Gentoo

[Anonymous Coward]

Especially if it's not thier main boxen.

boxen is not a word. It is a throwback to the old BSD days, when the developers needed an easy way to know if the person they were corresponding with was gay. Those that would say boxen were really saying they were promiscuously gay.

The purpose of the word has been lost, but it's use in the gay BSD community is now passe and is today used only by newbies.

HTH. Just trying to prevent you further embarassment.

Why I like Gentoo

[el_munkie]

I tried RH 6 or 7 a few years ago on my desktop machine, but I always had trouble installing programs. Dealing with dependencies, make files, etc. was completely beyond me, and I scrapped it and returned to windows.

Then I bought a mac and began playing around with fink. I had been using various Linux machines at school and had gotten pretty competent with the command line. I really liked the fact that fink checks dependencies, downloads the files, and installs them with one command. When I was showing this off to a friend, he told me about Gentoo and I decided to give it a shot.

I learned more about UNIX-type systems in the week-long install than I had in all my previous experience. The Gentoo installation process will give anyone with half a brain a fair idea how their system works by forcing them to deal with the various config files that they would not have to otherwise deal (until something messes up) in precompiled distros. I had never compiled a kernel before Gentoo, and they made that somewhat scary sounding task easy.

While the speed of programs compiled in portage is an important factor, my desktop seems much more responsive than it was under RH or Mandrake (two other distros I tested before my recent switch to Linux), it is not the selling point for me. It is the ease of installing programs that portage brings.

celeron

[asv108]

Celeron 2 GHz Processor 256 MB DDR RAM SAMSUNG - SP4002H 40G HD MSI 6533E main board All SIS chipset

I think the performance benefits are related to compiling for a specific arch say P4 over how standard distros package at i386 for compatability reasons. This test would be much more interesting if it was done with an Athlon XP or P4. I'm not familiar with celeron arch though, would /etc/make.conf setup be the same for a celeron as a p4? Why would anyone buy a celeron with Athlons at 2Ghz rated athlons going for peanuts?

Re:celeron

[Defiler]

Reasons why someone might buy a Celeron:

A) The Athlon XP 2000+ and Celeron 2GHz are within 5 dollars of each other on Pricewatch.

B) Most major vendors (including Dell) don't sell Athlons.

C) Celeron has a nice all-in-one chipset with Gigabit onboard, whereas the Athlon XP equivalent is significantly more expensive. (i.e. 8x5 chipsets vs. nForce2)

Re:right now

[Anonymous Coward]

Actually I use Debian (2.0.36 kernel), rh 6.2 (2 boxes), RH8 (1 smp box), RH7.3 (1 box), and Gentoo.

I just started using Gentoo. It rocks.

No distro is the king for "developers"

[0x0d0a]

Yeah...my PII/266 definitely takes less than 24 hrs to build X.

That being said, I'm dubious that blowing the time on compiling your ftp server with all optimizations every time you download a new version really is a worthwhile use of time and effort.

Maybe xmame. Maybe glibc. Maybe the kernel. That's about it. Definitely not 99% of the software on the system.

I don't really think any one distro is much better than the others for development. I happen to use Red Hat, which I do plenty of development on. I have a friend that uses SuSE and Mandrake, and another that uses Gentoo and Debian. All of us write software pretty happily -- all the tools we need are packaged or easy to build for whatever system we want to use. Most folks that I've seen that dislike a particular distro just plain don't know how to use the tools on that distro. I've liked just about every one I've tried, though they all have little things that one does better than the others -- RH shouldn't have shipped gcc 2.96, SuSE should put version numbers in their RPM names (actually, I believe they do these days), SuperRescue should be updated more often.

Heck, the vast majority of Linux users, not so long ago, *were* developers. By that metric, SuSE, Debian, and RH would be the most favored developer distros, though I doubt that contains one iota of useful information. :-)

Portage

[7-Vodka]

IMHO, the best thing about gentoo is the ease with which you can write an ebuild. This leads to a lot of software being supported and current. It's awesome. Have a small software package that's not in portage? spend 5 minutes on the ebuild and everyone can have it.
A new version of some software was freshly released? sometimes all it takes is renaming the ebuild. I hope they don't complicate this anymore than it already is; it's their greatest asset.

The question is: is this ease of packaging because the packages are mostly sources and not binaries?

Its about freedom!

[qoquaq]

One has the freedom to configure their system as they would like. To configure the system correctly one learns from going through the process. For those who are aspiring software developers, this is a rewarding challenge. Here is an opportunity to do this with a relatively easy system. For those with older workstations this is a better way to go such that you are in control of what to add to your system and how to optimize the kernel compile. It may not be twice as fast but its *your* system. Freedoms such as this should be highly valued and praised in todays software industry. Bravo gentoo!

Default Install = Stage1?

[aSiTiC]

I didn't see anywhere in the story if the Gentoo installation was done from scratch stage1 or from stage3. I would think this would be a very important piece of information to mention.

Re:Default Install = Stage1?

[Anonymous Coward]

True, that would have been interesting. If they are going to test if the source compiling and optimization does make Gentoo faster,... then they should have done the whole ball of wax.

Still, I didn't migrate to Gentoo from Redhat,Mandrake and Debian for the speed actually. It was purely: What distro do I feel confident enough on stability to run my finances, with packages like GnuCash and provide up to date applications.

Quicken was my last reason to use Windows with any regularity. Gentoo was the first distro that seemed to give me the stability I needed, coupled with newer apps. Debian is great on the stability also, but lot of the apps were older and lacking in some of the newer features I wanted.

Another good thing about Gentoo, is the forums. I haven't seen anything like it on another distro. People are much more helpful. Mandrake had something close, MandrakeUser.org. Not sure that was actually from Mandrake though, think it was a forum that started and they just linked to it.

Re:right now

[Anonymous Coward]

no kidding... what a dipshiat

There is no Gentoo

Gentoo sucks!!!!!!!!!

[Anonymous Coward]

Oh man, I tried this screwed up distro.

I didn't want to spend days compiling things like X, KDE, and GNOME so I opted for a couple of their binary installs (they obviously realized that compiling from source is stupid). What a joke, nothing worked, I couldn't even complete the install.

Whatever... Gentoo users are retarted morons.

There is no Gentoo

The Things that Really Matter in a distro

[0x0d0a]

It's not worth it. Moving from distro to distro for performance is pretty ridiculous.

Here's what I'd consider, since this is where the biggest differences lie:

* How frequently are new releases announced? Frequent new releases may be better for hobbyists, but a pain in the ass for servers sitting in a back room somewhere. (It's the reason RH can see an enterprise edition that's simply not released as frequently).

* How do you like the packaging system? Try out apt, emerge, up2date (actually, don't -- up2date truly sucks. Everyone using RH who cares about automatic updating has long since started using apt or (IMHO, better) yum).

* How do you like the config system? Most vendors have their own interface to let you configure the system. RH used to use linuxconf, and is now using Redhat-config. SuSE uses yast.

* How much do you care about commercial support? A few widely used distros tend to get the only commercial support. Mandrake gets a little, but if you're going to be running packages that require support (especially binary-only), you're probably best off with Red Hat.

* Which desktop environment do you want to use? Mandrake puts more work into KDE on their system, Red Hat into GNOME.

Arguments about speed or features is really pretty meaningless -- common software is generally packaged for most of these, and rare software for none (use checkinstall to *make* packages -- you'll be much happier). It's still Linux with the GNU suite present.

People that switch from distro to distro (or maintain *multiple* distros on their machine) are nuts, IMHO. It's a fair amount of work to relearn the quirks of each

Re:The Things that Really Matter in a distro

[Teach]

Try out apt, emerge, up2date (actually, don't -- up2date truly sucks. Everyone using RH who cares about automatic updating has long since started using apt or (IMHO, better) yum).

What don't you like about up2date? I think it's great, but then, I've never used apt or emerge, so I don't appreciate what you're comparing it to.

I keep the Red Hat Network Alert Notification Tool running in the tray, and then if it indicates that something's out-of-date, I just do

• su -c "up2date --nox --update"

...and BAM, everything's downloaded and installed.

Maybe you mean up2date sucks for installing new packages, rather than upgrading old ones? I get the feeling it's not really designed to do the same things as apt.

Re:Gentoo loses! (-O2 vs -O3)

[marcovje]

-O3 are space/speed tradeoff's.

The author's remark about -O2 for Celeron's probably hints at the smaller cache of Celeron's.

Space/speed tradeoffs can work to the disadvantage if the cache is small/bad. (because e.g. a loop isn't entirely cached)

However I doubt this goes for newer P4 based Celerons, since Celeron had their cache upped too through the years. So at least for code, they are
probably still better off with -O3 I think.
(which inlines small functions and does loop unrolling)

Has nobody posted this yet?

[Anonymous Coward]

Official Gentoo-Linux-Zealot translator-o-matic

Gentoo Linux is an interesting new distribution with some great features. Unfortunately, it has attracted a large number of clueless wannabes and leprotards who absolutely MUST advocate Gentoo at every opportunity. Let's look at the language of these zealots, and find out what it really means...

"Gentoo makes me so much more productive."
"Although I can't use the box at the moment because it's compiling something, as it will be for the next five days, it gives me more time to check out the latest USE flags and potentially unstable optimisation settings."

"Gentoo is more in the spirit of open source!"
"Apart from Hello World in Pascal at school, I've never written a single program in my life or contributed to an open source project, yet staring at endless streams of GCC output whizzing by somehow helps me contribute to international freedom."

"I use Gentoo because it's more like the BSDs."
"Last month I tried to install FreeBSD on a well-supported machine, but the text-based installer scared me off. I've never used a BSD, but the guys on Slashdot say that it's l33t though, so surely I must be for using Gentoo."

"Heh, my system is soooo much faster after installing Gentoo."
"I've spent hours recompiling Fetchmail, X-Chat, gEdit and thousands of other programs which spend 99% of their time waiting for user input. Even though only the kernel and glibc make a significant difference with optimisations, and RPMs and .debs can be rebuilt with a handful of commands (AND Red Hat supplies i686 kernel and glibc packages), my box MUST be faster. It's nothing to do with the fact that I've disabled all startup services and I'm running BlackBox instead of GNOME or KDE."

"...my Gentoo Linux workstation..."
"...my overclocked AMD eMachines box from PC World, and apart from the third-grade made-to-break components and dodgy fan..."

"You Red Hat guys must get sick of dependency hell..."
"I'm too stupid to understand that circular dependencies can be resolved by specifying BOTH .rpms together on the command line, and that problems hardly ever occur if one uses proper Red Hat packages instead of mixing SuSE, Mandrake and Joe's Linux packages together (which the system wasn't designed for)."

"All the other distros are soooo out of date."
"Constantly upgrading to the latest bleeding-edge untested software makes me more productive. Never mind the extensive testing and patching that Debian and Red Hat perform on their packages; I've just emerged the latest GNOME beta snapshot and compiled with -O9 -fomit-instructions, and it only crashes once every few hours."

"Let's face it, Gentoo is the future."
"OK, so no serious business is going to even consider Gentoo in the near future, and even with proper support and QA in place, it'll still eat up far too much of a company's valuable time. But this guy I met on #animepr0n is now using it, so it must be growing!"

-

Those cflags are nowhere near optimized

[SiliconJesus101]

Looks to me like they compiled using pretty much generic flags. As an exaple, my cflags are as follows: CFLAGS="-march=pentium4 -O3 -pipe -fomit-frame-pointer -fprefetch-loop-arrays -falign-functions=4 -funroll-loops -ffast-math -fforce-addr -mmmx -msse -msse2 -mfpmath=sse,387" Also, we may have to question their ability to properly compile a kernel. Possibly they did a generic kernel with everything but the kitchen sink thrown in. The idea of Gentoo is that you compile the kernel for YOUR system and not generically.

Re:Those cflags are nowhere near optimized

[Anonymous Coward]

Yes it is true that the longer your cflags are, the faster it actually runs code. Amazing!

Re:Those cflags are nowhere near optimized

[Sir-Tez]

You are absolutely right. Whoever did the gentoo install isn't properly informed on how to optimize it. I hope next time they borrow a gentoo dev and have them use the same hardware and THEN run the test.

I beg to differ subjectively

[marienf]

Although I never actually *measured* anything, I have been moving all my boxen (except for one Duron on which I have found it quite impossible to compile Gentoo) to Gentoo 1.4rc4. I was actually in the process of building my own compile-in-place GNU/Linux called "Q-Gnu/Linux" when I discovered Gentoo did it all, and did it better. I was all RedHat before that (going so far as to wear a red fedora on parties - I have two of those). I find Gentoo as opposed to RedHat quite impressive, at least. My professional workhorse (on which I'm currently typing) is a Toshiba Satellite Pro 4300:

model name : Celeron (Coppermine)
stepping : 3
cpu MHz : 597.077
cache size : 128 KB

..with 384MB RAM.. and was becoming annoyingly slow in things requiring major GUI complexity, like OpenOffice, and at compiling many Java classes.

Compiling Gentoo on there allowed the machine a third chance at life, the second one being when I got it (already old then) and installed RedHat on it, over that would-be-OS it came with. It just feels that much faster again. I am no longer annoyed by it at all. It took more than 4 days to compile all I wanted from the Gentoo 1.4rc4, but it was *well* worth it.

I moved my personal little server, an Athlon Thunderbird, with the same impression. Currently running

emerge system

on my brand new Athlon XP 2600, expecting much from it.

Bottom line: Nothing but Kudos for Gentoo, wondering what went wrong during the tests described, or whether somehow the subjective speedups I have experienced are just auto-suggestion. I think not. I have been staring at CRT's since 1980, thats 23 years folks! And I tell ya compiling stuff yourself is worth it. So if you have time on your side, go for LFS, which I did, and slowly ground into Q-GNU/Linux. If you have some time, but not *that* much time, go for Gentoo, if you have no time, you poor shmuck, either get a life, or install SuSe :-), and pretend.. :-) :-)..

Re:right now

[0x0d0a]

If you invest a lot of time in learning a distro, you're terrified that it might not be the best, and will spend ridiculous amounts of time insulting the others.

Hence, the distro wars.

Re:Remember the KDE mandrake/gentoo fiasco?

[buchanmilne]

Mandrake while my favorite choice, doesnt include the best pre-emptive kernels.

You mean like this one (from contrib for 9.1)?


Name : kernel-multimedia-2.4.21.0.16mdk
Group : System/Kernel and hardware Source RPM: kernel-multimedia-2.4.21.0.16mdk-1-1mdk.src.rpm
L icense: GPL
Packager : Danny Tholen
URL : http://www.kernel.org/
Summary : A preemptible Linux kernel, which reduces the latency of the kernel.
Description :
This kernel includes patches useful for multmedia purposes like:
preemption, low-latency and the ability for processes to transfer their
capabilities.
The preemtion patches allow a task to be preempted anywhere within the kernel,
using spinlocks as markers for non-preemptibility regions. The resulting
system response is greatly increased, with measured average latencies under
1ms. Andrew Morton's low-latency patches fix the remaining points in the kernel
that cause latency. The setpcap patch allows suid root processess to transfer
capabilities to non-root processess, and so making it possible for user
processes to run with realtime priority.


[some uninteresting fields removed in aid of the lameness filter]

The next one for 9.2 contrib will most likely have the O(1) scheduler also.

This article is flawed...

[eWarz]

They optimized Gentoo for the p3 platform? Celeron 1.4 ghz and above is based on the p4 core.

portage does do binary packages.

[Anonymous Coward]

maybe you can read up on portage to find out?

(-k) Tells emerge to use binary packages (from $PKGDIR) if they are available, thus possibly avoiding some time-consuming compiles. This option is useful for CD installs; you can export PKGDIR=/mnt/cdrom/packages and then use this option to have emerge "pull" binary packages from the CD in order to satisfy dependencies.

Re:Remember the KDE mandrake/gentoo fiasco?

[trashme]

Myself, Gentoo's biggest feature was the kernal compile options, adding patches for pre-emptive mulitasking, and improved responsiveness.

Ahem. I'm guessing you are talking about the preemptive kernel patch. Linux, and every other modern OS, already have preemptive multi-tasking. Preempting the kernel is different from preemptive multi-tasking.

Preemptive multi-tasking just means that a process can be interrupted at any point and another process, or the OS, can make use of the CPU. Preempting the kernel means that actual kernel code can be interrupted.

There are other distributions that include kernel patches as packages. Debian, for example, let's you patch your kernel with a compile option passed to make-kpkg.

Linux is Linux

[lurid980]

I really don't understand how Linux is supposed to run faster just because one group of people, as opposed to another, package some OSS software. Its the same software! Who cares if its Red Hat or Debian, gcc is gcc. If you're testing 3 distros on the same hardware, with the same software, written by the same people.. how is it that we honestly expect one to be faster than the other?

Its not like we're comparing Windows 95 with Windows XP, where the entire underlying libraries are vastly different.. we're talking about Debian distributing GCC and Gentoo distributing the same thing. While compiling for an XP processor only vs compiling for anything made since the 386 will yield performance gains, none of them are truely 'real-world' gains.

I use Gentoo myself. I enjoy staying bleeding edge. I do not run a production server so the very "ancient, yet stable" mantra of Debian does not interest me. I'm running 2.6.0-test2-mm2 for my kernel right now. I like playing with software. Gentoo is aimed at a person like me. I can tweak it, play with it and fully enjoy my computer and my OS. Thats not for you? Great! We have about 3 million other distros, all with their own aims, for you to choose from.

Honestly I think this in-fighting in the Linux community is sickening. Are we really a community when we trash ourselves, who are all on the same team, as much as we trash the competition? I like and use Gentoo. Its made for people like me. Rather than trash Gentoo about long compile times, I believe that you should simply identify what you would like to benefit from Linux and pick a distribution of it that suites your needs. We need to act like a community again, not like spoiled brats.

Re:Linux is Linux

[erikharrison]

While compiling for an XP processor only vs compiling for anything made since the 386 will yield performance gains, none of them are truely 'real-world' gains.

Then why were the archetecture changes made in the first place?

Re:Linux is Linux

[lurid980]

The point is that if you have an XP processor and you compile something for i386, i686 or -march-athlon-xp you're not going to see major speedups. The point is that installing via RPM or compiling with massive Gentoo tweaks.. you're using the same software. Why we all fight over whos the best/fastest I'll never understand. We all use Linux.

Re:Linux is Linux

[Anonymous Coward]

Not by your pervious comments, you seem to say that you like and use XP. You sound like a little troll from Microsoft, we've been getting a lot of those lately.

As for your post, there is a difference when setting up the system, RPM is probably the worst package management system. Mostly because of the bugs, and the lack of a clear scheme of things.

As your your comment on i386 to i686, why not try a system that has been optimized for the latter. You might notice a difference. The actual package will matter less, but as a system whole it will.

Re:Remember the KDE mandrake/gentoo fiasco?

[leviramsey]

For a litle while, Danny Tholen (one of the helpers on the mdk kernel) had a multimedia kernel in Cooker contribs, using backported preempt and so forth patches.

cflags and different sources per distro

[Anonymous Coward]

I think they should make sure to use the same cflags as debian or mandrake, except the -march setting. Last I did a src.rpm compile on a mandrake distro they used flags like -ffast-math which makes a big difference. I've also seen people using stupid flags like -mfpmath=sse or -mfpmath=sse,387 which makes performance worse. Some people tend to add both -march and -mcpu which is silly. There's lots of things you can do wrong with cflags. Also the distros are not compiled from the same source so it's not fair at all. Every distro has it's own patches and other things they do to optimize.

I don't think this is proof that compiling for your cpu doesn't help, since there's so many other things that can impact performance. If they did the same test with _all_ gentoo systems, then at least one would know whether cflags matter much or not.

This should be interesting for SOME people

[defile]

One of the arguments I've always heard incorrectly used to promote Gentoo is that your system will BE faster because you've built everything from source with optimizations enabled.

I think this is largely BS.

In modern computing environments, the system bottleneck is not commonly the CPU. Most workstations and many servers, as they're used today, simply spend all of their time waiting for something interesting to happen. Something interesting usually means waiting for the user to push a button, waiting for the disk to send a chunk of data back, waiting for some packet to take a trip back from over the network. More MIPS and nothing else simply doesn't make a difference for the common workload.

This doesn't mean that there aren't cases where a system will spend a lot of time burning on the CPU. But these cases are often identified by their developers and are addressed accordingly, in a variety of ways. The developer may choose to employ a more complex but faster algorithm, offload it to specialized hardware, rewrite in a lower level language, etc.

If the package you're installing doesn't have compiler optimizations already set in the Makefile, there's probably a reason for it.

Therefore, blindly taking each package and running them through gcc -fomit-frame-pointer -O8 -march=i686 -mcpu=i686 blah blah is basically system admin masturbation.

I like Gentoo, I'm running it right now, but I don't use the optimizations argument.

Re:This should be interesting for SOME people

[dh003i]

Different packages need different optimizations. Packages that have a huge mess of stuff to be loaded probably need -Os, not -O2 or -O3. Not all packages benefit from fomit-frame-pointer. Some packages -- e.g,. computationally intensive scientific packages -- will benefit significantly from CPU-optimizations. Others won't.

To take advantage of Gentoo's optimizations, you need to understand what various optimizations do and what your packages need.

Re:Remember the KDE mandrake/gentoo fiasco?

[Anonymous Coward]

30% Redundant

Really wish the moderators would look at post times. Its not redundant if posted first.

Moore's Law

[nagora]

The time to install from source halves every 18 months. Already, entry-level systems can compile a Gentoo desktop system up from stage1 to everything-except-OO in a day (and OO can be installed as a binary) and a server can be up and running in a couple of hours so compile time is not a big deal and gets less so every day.

Given how much better Portage is than any of the other management systems, I'd say Redhat is going to suffer big loses at the hands of Gentoo (Debian would too but the effect will be drowned out by the damage Debian is doing to Debian).

So far I've converted six machines to Gentoo, all from Redhat because I couldn't face upgrading with RPMs anymore.

TWW

Re:Moore's Law

[sholden]

For modern bloated software compile times are influenced by disk speed as well as CPU speed. Disk speed does not double every 18 months.

Also every 18 months software bloats some more, slowing down compile times again.

And I don't buy a new CPU every 18 months, in fact I the newest CPU in my home machines is 5 years old.

Compiling everything myself would inconveniance my way of using linux machines, especially fresh installs.

Essentially I install the base debian system, dist-upgrade to unstable, install X, and then use the system. For the fist week or so I do "sudo apt-get install foo" often as I discover a package I want but don't have yet (well the first time I try that I do "apt-get install sudo" as root after it fails, of course).

Just in time installation is wonderful, the system isn't too bloated by lots of software I don't use. My style of system installation won't work well with gentoo because installation of a package will take more than 10 seconds, and source downloads are significantly bigger than binary downloads slowing it further (of course my method is only useful for desktop machines too).

WIth 5 year old machine, you may wonder just how often I do a fresh install, and the answer is not often at home, but at uni it seems to be a frequent task.

I write perl and C++ code, obviously squeezing that extra bit of speed out of the hardware just isn't a priotity to me :)

For people for whom it is, obviously gentoo can make a faster machine that a one-size-fits-many binary distribution.

Re:Moore's Law

[Anonymous Coward]

Using moore's law as an argument for Gentoo is a double-edged sword my friend.

Why bother spending time compiling stuff yourself for more speed, when new processors are coming out seemingly all the time that are faster and faster for the same price? But do you really get more speed? From what I can gather from reading people's coments - the hardcore coders have been saying that these compiler optimisations make very little difference (expcept for 586 optimisations on the original Pentium).

It seems like the real reason people why people who install Gentoo and say "hey i've installed Gentoo and I can feel it's faster!" is because of the heavy patches they put into the kernel for interactive use. All these interactive improvements are coming with the 2.6 kernel anyways so the binary distros just have to wait another 6 months or so, and I believe the whole "gentoo feels faster" argument will be mute. Besides, my Athlon XP2000+ feels fine for interactive use! Also, I'd rather use a kernel from an official kernel.org mirror for the sake of compatibility. ie how do I know that these interactive patches are causing progam 'x' to crash? Stock 2.6 kernel will get much more testing that gentoo's custom heavily-patched kernel. In fact the sheer number of patches that go into this kernel, quite frankly, scares me! (from a stability/testing standpoint).

Speed arguments be damned - what it all boils down to is ease of management. How do I get what I want onto my system in the easiest possible way? (most uptodate stuff, proprietary mp3 stuff etc). Also how do i get packages I don't like, or were merely testing, off my machine without stuffing things up? How do I best avoid dependancy hell? I think the distro who solves package management issues the best wins. Now is this going to be deb (with rpm), rpm (possibly with rpm-apt) or portage?

celeron 2ghz is pentium4 arch not pentium3

[Technomancer]

there are couple more optimization flags that would have helped.
but thats not the only problem. they used different display drivers and it could be possible that framebuffer driver is better than sis driver (especially that this looks like shared framebuffer system).

Re:celeron 2ghz is pentium4 arch not pentium3

[Anonymous Coward]

No sorry. The framebuffer is NEVER faster than any accelerated driver. Even the SiS integrated chipsets (hell, the ones from ten years ago), whether shared memory or not, have hardware acceleration, and this won't be used by a framebuffer driver.

Re:No. Gentoo is not a performance leader.

[Squinky86]

I know I use it for the speed gain, and to say that compiling your own packages doesn't matter is absurd! If you know how to and properly configure your compiling flags, the speed gains are tremendous. I also agree completely that disk i/o and fs type are the determining factors for many tests, so I agree with the statements of programs running at the same speed while writting from or reading to disks, but if you just try gentoo, the learning experience and speed gains are very noticeable.

My *nix usage

[BenjyD]

In the last few years, I've tried:

RedHat6.2 (what's this whole linucks thing about?)
Solaris 8 intel (yuck - slow, hard to use)
SuSE 7.0 (ick)
SuSE 7.3 (I think, still ick, crashed)
Debian Woody (Outdated but good. Got fed up with old versions of everything eventually)
FreeBSD 4.8 (Fast+nice, no sound, barfed on my USB mouse for some reason occasionally though. Gave up fiddling)
SuSE 8.2 (Why do I go back to SuSE? Still sucks. What is this awful YAST thing doing?)
Gentoo 1.4rc-whatever-they're-up-to-now (stable, fast up-to-date. Takes forever to install)

Personally, I'm sticking with Gentoo. It just works. type emerge xine and half-an-hour later I'm watching the Matrix2 trailer.

CFLAGS detection script

[cosjef]

Here's a script that helps you determines which CFLAGS are compatible with your CPU:

Compiler options: sometimes matters, usually not

[Ogerman]

From my personal experience, I can attest to the validity of the article's argument. It has been my experience that general purpose software (X11,KDE,etc.) does not gain any significant performance advantage when compiled from source with tweaked gcc compiler options. And there are plenty of other folks who have for years claimed the same thing. Perhaps the most significant reason for this is that modern x86 processors do their own optimizations (instruction re-ordering, etc.) internally. So the vanilla "i386" option is just fine in most cases. The other reason why option tweaking is pointless is that developers typically set their own optimized compiler settings in their Makefiles if it's really going to make that much of a difference. (and incidentally, some compiler optimizations such as -ffast-math aren't even safe to apply generically to all software on a machine) So this whole "which distro is faster" thing really is a dead issue.

That being said, there are a handful of cases where optimizing for your specific processor CAN slightly improve performance. (typically heavy number crunching routines). Some software like gzip, bzip2, gnupg, povray, fftw or other math libraries, etc. may benefit. But that's no need for Gentoo. Debian source packages are trivially easy to retrieve, set optimizations, and compile. It's highly automated, just like Gentoo. The only difference is you aren't forced to do that for every piece of software on your machine.

Loop unrolling

[Anonymous Coward]

I don't think -O3 does loop unrolling, based on the GCC docs. Inlining, yes.

Re:Loop unrolling

[Anonymous Coward]

-funroll-loops

It's the uninstall that bugs me.

[cynicalcdb]

Gentoo certainly isn't a /bad/ distro, imo. In fact the portage system is quite nice for installing software. It's when you want to uninstall something that you run into problems. AFAIK, emerge/portage still has no ability to warn you if you're about to unmerge something that is depend for something else installed. Nor is there a quick, easy way to find out which of your /installed/ packages depend on which other installed packages. That, I think, is it's biggest shortcoming.

Re:right now

[Anonymous Coward]

If you invest a lot of time in learning a distro, you're terrified that it might not be the best, and will spend ridiculous amounts of time insulting the others.

Maybe we should start a sort of parody, where we pretend to be massively loyal to different desktop environments and each constantly flame the other one. It might show these distro-warriors how silly they're being.

The kernel compile is perfectly fair

[El Cubano]

Also, the kernel compile is unfair, because gentoo-sources includes a whole load of patches that Mandrake and Debian don't.

From the article:

The same 2.4.21 source was copied to all machines and compiled using the same options.

That looks pretty fair to me.

Re:The kernel compile is perfectly fair

[Anonymous Coward]

But it misses the point of having a tweaked distribution. Of course the distributions are going to be the same if they all run identical software.

Re:Compiler options: sometimes matters, usually no

[digitalhermit]

Completely agree. I tried Gentoo for a while, emerging an entire X11 system with KDE from stage 1. The gains I noticed turned out to be because of the improved KDE, not from the optimizations. If there was any gain in speed it was more than offset by the time it took to bring in new updates and troubleshoot the ones that didn't work correctly.

As you also mentioned, I did notice a small increase (maybe 5-10%) after recompiling my povray binaries with aggressive optimizations. I did get a vast improvement after using a demo version of the Intel compilers and a pretty good improvement after linking with specialized math libraries. Of course, these gains were also offset by the time it took to implement them. And considering that a $50 CPU upgrade gave a twofold improvement, I'm more likely to upgrade hardware now than spend time tweaking.

To be fair, I should mention that I started using more lightweight window managers (fluxbox) and made an ancient laptop usable once again.

The benchmark doesn't compile

[GGardner]

I notice that your benchmark doesn't compile. I'm curious where your rand.h came from, and what

float something = rand[loop] + rand[loop + 1];

is supposed to do. rand is an ANSI C function.

Just wondering...

[aLEczapKA]

First of all, I switched from Slack (tried Debian and SuSe before) to Gentoo and I did noticed the difference in speed. In real life, Gentoo is faster!

Second, don't you think there is something wrong with this test? Did I only noticed that or what? Please somebody explain me how is it possible that Mandrake won almost all of the tests?
Mandrake was compiled for i586, and it was runing on i686 so the logical conlclusion would be to compile everything as i586 but to run it on i686..? Something definitly wrong here...

I have fast machine (amd 1800+) and I don't care about compiling time, besides it goes fast, so why bother?
Why do you complain that it takes so looong to compile stuff? Ofkoz if you have old machine - but you don't have to - you can always get binaries.. so what's the deal??

This test is wrong and prooves nothing - I run Gentoo and I can see it's faster.
It just requires little time to read about how to tweak it and than just to compile what you need.

You could say Gentoo is like LFS but with very good packaging system (this is what LFS lacks) and it's much more easy to manage.

Re:Just wondering...

[Anonymous Coward]

In real life, Gentoo is faster!

Sorry, thats meaningless. It doesn't matter how it feels to you. There are many, many things affecting that. When you get a Gentoo system into a usable state (ie, the userland, kernel, and boot loader are functional, so you can reboot the system and install software) it has basically nothing installed. You manually install every single thing you want, and there are next to no services running. Now even with slackware, if you check one of the default install options, it will install a lot of server software that strictly speaking you probably don't need, and all of that will be running at startup. And yeah, it might feel faster without that stuff running. Has nothing whatsoever to do with Gentoo. Also their kernel has all sorts of patches for preemption, etc. applied which Slackware will not touch until they are well-tested and part of the default kernel. That will make it seem faster too, but has nothing to do with Gentoo per se.

Can you see that this test has a level of objectivity beyond your personal experiences? Each distribution was installed on the same hardware and running the same kernel, and everything else was in a bone-stock configuration. They ran these tests, and those numbers are what you see. Sure you will whinge, because it invalidates the several days you spent building your entire system from source, but if you want to complain about this test you are going to have to come up with a superior methodology which produces different results, not just personal anecdotes.


You could say Gentoo is like LFS but with very good packaging system (this is what LFS lacks) and it's much more easy to manage.

No. Its not "from scratch" at all in the sense that all meanigful steps in the installation are automated, and the remainder are formulaic enough that you can understand why most distributions script the installation. The benefit of LFS is that, by installing each piece yourself and attending to every bit of configuration, you gain a deeper understanding of how the system works. You get next to none of that from Gentoo, except the massive time requirement.

Re:Just wondering...

[aLEczapKA]

It doesn't matter how it feels to you
If it wouldn't matter I wouldn't use it, right? And I can _feel_ it the way that I can see how fast several applications start, to name some: mozilla, nautilus, window managers - they start faster, I can see it therefore I feel it, that's what I meant.

The things you menshion about slackware - as you noticed - they are Gentoo related so they have everything to do with Gentoo. Starting from the point when applications, I don't need, do not start and ending up on the kernel patches. Btw: my kernel works very well, and I didn't notice any crushes at all since I use it.

Slackware has just different approach - security first, but hey this is my desktop machine it's not online all the time, no other users here, I don't have to be paranoid about security and stability, it could crush from time to time, I don't care so I do use Gentoo kernel patches, which as I said, didn't cause me any troubles yet.

Its not "from scratch"..
Yes, it is in terms you compile stuff from the begining including gcc and such... optimized for you machine...

The benefit of LFS is that ... ou gain a deeper understanding of how the system works

You're right but it also depends what you are expecting, you learn - that for sure, but in the end you have optimized, smaller (just the stuff you really need) and faster system.

I am not whining about the test just wondering, as I wrote already why mandrake is faster even if it was compiled for i586 but the test ran on i686???
Just wanted to point out that guys doing the test for sure did someting wrong.

And don't worry about my installation and time I spend on it - one test will not make me change the distro dude!

Re:Just wondering...

[aLEczapKA]

One more thing. Just measured that.

Gentoo boots in 25 seconds (to the shell, and I do start some stuff, sshd, postifx, saslauth, etc).

6 seconds later I have X running with Fluxbox (with gdm & autologin) + bunch of dockapps.

So it's like 31 seconds all toghether. This is very good result as for the desktop system. Bit that! Than we can talk.

ICC

[Ann Coulter]

Has anyone tried to use Intel's icc compiler to compile Gentoo?

Re:ICC

[Anonymous Coward]

It's something that is being worked on. Hopefully there will be an install profile soon that uses predominantly icc. Check #gentoo-icc irc channels on freenode and oftc.

Tried Gentoo once...

[HouseOfMisterE]

I first jumped onto the linux bandwagon with Redhat 6.0, and have tried several different distros since (Mandrake, Suse, etc.) Tried Gentoo once. After spending over a day trying to install and compile (and troubleshoot), the damn thing wouldn't work. No boot for me. Said "fuk it," and went back to Redhat. Having spent so many years around non-optimized operating systems (i.e. Microsoft Windows), I found that the Gentoo headache wasn't worth it. To each their own, though. E.

Source vs. JIT optimization

[simm_s]

I simply do not like distributions you have to compile because it is just too time consuming. I only compile software when the software is a source-only release or when I want to patch the software. Spending half a day compiling my distribution is unacceptible.

I like the .NET/CLR approach of compiling the code into an intermediate format and then running a just-in-time compiler during the first run. Only the code that is used is compilied (with specific optimizations for the machine) and the code is permenantly cached.

This solves three problems that source code compilation/optimation does not:
* its faster than lexing, compiling, and linking
* it allows me to optimize binary-only software
* it allows the intermediate binaries to be cross-platform.

This way the developers can spend time optimizing the code to a virtual target platform, and the administrators can spend time administrating. ;-)

So is linux

[r6144]

Especially when you link to a lot of libraries, dynamic linking can often take a lot of milliseconds. Prelinking helps a bit, but static linking is the fastest. On my machine running a simplest program takes 5ms when dynamically linked, 3ms when statically linked, "user time" is 1.1ms vs. 0.4ms.

Re:So is linux

[pantherace]

prelink can help with this. google for it and you should find out about it or look on gentoo.org under the docs section. Prelink support is built into portage :)

Gentoo makes me understand Linux better.

[cnmill]

I have bouced back between Mandrake and Redhat for a couple of years. Tried Gentoo last mo.nth, and although it has been quite challenging for me, I have a mich greater understanding now of how Linux works. It has forced me to go beyond the GUI (and the re-install if I really dorked things up.)

This is why I enjoy tikering wiht Linux... it's like working on cars without getting you hands dirty.

Re:Gentoo makes me understand Linux better.

[jimmy_dean]

Yes, I have had exactly the same experience. I am a better Linux user and in general a more knowledgable computer user in general as a result of using Gentoo Linux.

Working with debs.

[dmaxwell]

debs are actually renamed ar archives that conform to a specification and have been renamed. I have come across guides on unaring a deb,altering it and then packing it back up. I've never had to do it but the guides on how to do it didn't seem too bad.

I've never had a dependency problem I couldn't fix in Debian with a little noodling around with the system. On the other hand, I do sometimes recompile source debs to get options that I need. For instance, I deploy Netatalk with dhx authentication. I have to install the crypto devel libs and tweak the rules file to get it. Thankfully, I can build on one box and deploy the custom deb whereever needed.

I suspect that I could get most of the benefit of Gentoo by rebuilding the kernel, glibc, and maybe the xlib source debs with i686 options. At some point, I would like to see where the most benefit is reaped from custom compilation.

Re:No. Gentoo is not a performance leader.

[simm_s]

Do you have any evidence to back this up? Did you compile and benchmark a program with default optimizations and then compile and benchmark the same program with optimizations?

Why screw around with compiler flags when you can "crack open" the source code. I guaruntee if you are a competent hacker you can get more performance out of your software by doing that.

Although I agree with you that you may learn alot by compiling source yourself and learning how to optimize executable binaries, you will learn and possibly contribute even more by optimizing the source code.

Afford

[Mark_MF-WN]

You can afford all those electronic devices, but you still use a non-standard (crufty; that is to say, shitty) CPU? That seems a little silly. Pick yourself up a nice G5 or an Athlon 2600+. Stable as all hell, and speedy to boot. Once you've done that, be a real man and install Mandrake post-haste. You can use the time save from compiling to go and kill a wild animal or have sex with something. Or even just to play a rousing game of Pingus or Star Control.

Re:Afford

[Anonymous Coward]

It may be shitty but it is also very quiet. Something you can't say of a G5 or Athlon 2600

Re:Afford

[Anonymous Coward]

You have no idea what a Via C3 is used in, do you?

All this really shows is that these were bad tests

[oobar]

When I see things like the program time going from 39m 08s to 11m 21s (when all that was changed was a minor version number) that just screams -bad testing-.

You should repeat every one of the tests a number of times, and make sure that you get the same (or similar) results each time. You should not NEVER expect a 4:1 ratio of performace doing the exact same task on identical hardware. Bells should be going off that say "casual testing" when you see something like that.

Besides, there are so many variables that have to be kept the same between the different installs - which services are running, how they are configured, what kernel options are set, what patches have been applied to the kernel, which modules are loaded... If you pick up Redhat 9 and do a "kitchen sink" install, you will hardly have the same amount of free RAM for caching, etc. compared to doing the "regular" install of some other distro that leaves out things. Hopefully it's obvious that such a comparison that would not be fair at all.

In short, you should take a given kernel source, with a fixed set of patches, options, settings, modules, etc., and complile it with the default i386 options and then a second time with all the fancy optimizaions, then compare those. LEAVE EVERYTHING ELSE THE SAME! Repeat with glibc.

The results in this article are just pathetic. They vary all over the place and are crying out for more rigorous testing methods and procedures. Making a good test is really a science, you have to design the test to specifically measure what it is that you're interested in. For all we know one of those tests could have already had a majority of the libraries loaded into the disk cache, resulting in the huge performance differences.

And pentium 4

[r6144]

I have heard that P4 runs code compiled for i386/P2/P3 rather slowly compared to code compiled for it. For example, P4 runs traditional (stack-based) FPU instructions rather slowly, while it prefers SSE/SSE2-based instructions. Therefore on P4 systems compiling with the right options may well give a significant speed boost.

However, AFAIK PPro/P2/P3/Athlon runs these "legacy" code quite well, so relatively little gain can come from compiler option tweakings.

This *IS* the problem with Gentoo

[swordgeek]

Install Gentoo without knowing what you're doing, and surprise surprise--it's not necessarily better.

Gentoo is NOT a distro for everyone, folks! It takes work, understanding, and patience to get it working, and working optimally. If you don't have the understanding or patience, or don't want to put the work into it, then stick with Debian, or some lesser distro. :-) (OK OK, some _other_ distro--I at least admit to being a Debian bigot :-)

Seriously, Gentoo reminds me of SunOS kernel tweaking in the bad old days. Someone would find a stat that approximately related to a given performance claim. Then they'd tune the kernel to improve that stat, without realising that

a) They'd broken the relationship between the stat and the performance it was supposed to be measuring, and

b) They'd slowed down the rest of the system horribly.

I would strongly expect that Gentoo is faster (maybe significantly, i.e. on the realm of 10-25%) than other distros under certain circumstances, most notably under very careful install and setup. However, the 'generic' defaults that RedHat and others come up with to address most of the population just aren't that bad.

Re:This *IS* the problem with Gentoo

[lanalyst]

Well, we have to all start somewhere. And maybe for a few months there's a bunch of folks out there running with mediocre compiler flags.. and they're kernel isnt tweaked .. and so on. As time goes on, folks generally want to try optimizing and run a few benchmarks and observe results.. progress! Accomplishment!

For myself, I'd never claim my system is running optimally: there are constant updates (features and fixes) that make it almost impossible to keep up with it all... and that's great.

In the end, Gentoo makes better troubleshooters.. generally folks develop good habits: don't change a lot of things at once.. run benchmarks, measure differences.. UNDERSTAND what different kernel options do and how they impact the system overall.

It's folks that grab a LiveCD that have never run a source based distro - and write articles comparing Gentoo (or anything else) against a distro they have run for years that makes me chuckle. Reminds me of 'the windows guy' writing a RedHat review on NewsForge a few months ago. Pretty funny. And totally missed the point.

I'd like to see 3 systems side by side each built by seasoned users of each distro. I'd bet the Gentoo user would be able to give the other two a few tips to optimize their systems. The gentoo user is pretty savy about versions and releases. Afterall, if they run a package, they can receive and run every update.

Another thing that's interesting.. it's really discouraged to overclock a Gentoo System. Long compiles will segfault. It turns out to be more fun actually experimenting with software configurations ... learn something new that way...

A couple of discussion related questions

[Xua]

I am going to move to gentoo. The reasons are

- very last versions of packages (Mandrake that I use at the moment usually doesn't create new packages until the next distro version, KDE is an exception, but is very slow to appear). I know about www.pclinuxonline.com and Texstar's rpms are the reason why I didn't move from Mandrake yet, but I think I'll change to gentoo anyway.

- stock package versions - I don't like distro to change the behavior of how the whole system works, KDE is an example again, Mosfet wrote about Redhat and Mandrake and why he doesn't like them. I agree with him there.

- Compile everything for a used platform. Now here are the questions:

* Is it worth to really use SSE2 math when compiling for P4? I don't remember particular gcc options, but it can use SSE2 math library and SSE2 floating point arithmetics.

* Any other optimization options that people use often? I searched gentoo forum, everyone seems to stop after -fomit-frame-pointer. I know gcc can generate wrong code when strong optimizations are used, but any experience with it?

* Can gentoo use Intel compiler for building the system? I've seen tests and they had shown that Intel compiler generates better code even for Athlons (I have one at home).

- Portage so far seems to be comparable with urpmi in Mandrake. I didn't notice any advantages, but who knows. I was disappointed in Debian packager once, maybe there are hidden features in portage that I won't live without after I find them.

irrelevant to test situation

[Anonymous Coward]

Distcc is "distributed c(++) compiler", and as such only makes a difference if you can use 2 or more machines for the compile. Ccache- compiler cache- doesn't make any difference on the first time compiling. Therefore, these pieces of software being installed on a machine make absolutely no difference in themselves wrt a benchmark compile.

Re:irrelevant to test situation

[Anonymous Coward]

Wrong! Because of re-used code and duplicate headers, ccache does make a difference.

Re:No. Gentoo is not a performance leader.

[hoddi]

Once I set out to prove that wrong. My test was purely cpu bound, to see the real benfit from the omptimization. I have since lost my resaults ... but you can do it yourself ... Here is what I did time cat /dev/kcore | gzip -f > /dev/null My resaults showed that whith correct optimization I would gain upto 10% increse in throughput. But, in the real world, the gain would be less.

Re:No. Gentoo is not a performance leader.

[Squinky86]

yeah, actually I did, but the results are lengthy and boring. I just say that a simple O3 and arch=(cpu) do make a big difference.
I love working with source code, too. Many of my programs have custom parts in them, and I also write my own. I agree wholeheartedly with all that you said.

Bloat

[rendler]

That's the only advantage I can see to doing things the Gentoo way. Someone mentioned this on here awhile ago and I really didn't get what he was talking about until he elaborated on it a bit more to me. What it comes down to is why do I need to install libldap just because exim is linked to it. I don't use any LDAP features in exim so why must I be FORCED to have it on my system. Same goes for a lot of other packages out there. With every dist-upgrade/upgrade (well almost every) there is some new library that needs to be installed because a package(s) already installed has had an option turned on that requires extra libraries. Of course with the Gentoo way when you're compiling your own packages you can turn those options off and cut down on the bloat quite easily rather than having to jump through hoops rebuilding a customized deb.

Define 'same'

[The Monster]

Except in this case they all had the same hardware on each machine...

That is frankly impossible. Even though the machines were supposed to be identical:

Upon testing with hdparm, it was apparent that this machine was having troubles setting above udma2. Eventually this problem was traced to the HD cable, a salutary lesson in the variability of identical hardware setups.

This is just the difference they caught. Who knows how many other subltle variations exist between nominally identical machines? An honest attempt to determine how fast 3 distros do the same thing would be to really use the same hardware, by running the tests on one or more machines with one distro, then wiping the HDs and installing the second and repeating the tests, then on to the third.

The only way to have the same hardware is to use the same machine for each distro. Period.

Re:Define 'same'

[dhawton]

I agree here. Except for the fact that between installations something might happen to the hardware, ie, an IDE cable became bent or the hard drive got too hot and melted, etc. That would suck. :)

Re:Define 'same'

[Anonymous Coward]

Hogwash. Granted, identical hardware is not the same hardware, but I defy anyone to be able to measure a difference.
A dented or whatever IDE cable is an anomaly, a defect, having it only affect performance and not cause spectacular crashes is a remarkable coincidence. It is unlikely that there would be say, another dent in a cable somewhere that only affected performance and not the actual functionality of the machine.
CPU's, memory, graphic cards, NICs and motherboards of the same designation are for all measurable purposes identical, which is why you can have SMP machines and mix memory from different manufacturers for example.
If you wipe the drive and reinstall on the same unit, one could still argue that "one distro got it's swap file alloated in a more optimal way, no fair!"

Re:Define 'same'

[VPN3000]

Agreed. Back when I was a hardware tech in the early 90's, I recall building pools of identical machines for customer orders. For burn-in, I would loop benchmarks for 24 hours before shipping them out. There was typically 1-2% difference in identical systems.

People tend to forget the complexity of a PC and the inevitable, microscopic differences each part made. Thus differences in resistance, heat generated, and performance.

BSD Ports...

[nurb432]

Not to sound like a troll.. But these sorts of questions have been answered long ago by the BSD crowd..

Its still up to the individual, but its been shown going 'make world' type of route is the best choice, if you can do it.

nothing to see here

[Anonymous Coward]

Linux geek believes sample size of 1 can yield "results". News at 11.

In order for these numbers to even be looked at, the tests would have to be run a couple dozen times each at least.

Hrmm

[Anonymous Coward]

My system runs gentoo and i just switched off from mandrake 9.1... I've noticed some performance boosts but like others have said i didn't change for the speed i did it for portage i love that little tool... Also i run -O3 and a bunch of other flags I have a 1.8GHz celeron laptop or as a like to call them p4lite... They should rerun the test with -03 and other commonly used flags... I'm sure most gentoo users compile wiht more then a two or 3 flags... they need to list there hdparm settings also...

I ran "-march=pentium4 -O3 -mmmx -msse -msse2 -mfpmath=sse,387 -pipe -fomit-frame-pointer -funroll-loops -funroll-all-loops -fthread-jumps -frerun-cse-after-loop -frerun-loop-opt -fexpensive-optimizations -falign-functions=4 -falign-jumps=4" for a month wiht heavy usage and i had no problems... then i had to change back to windows... stupid job and stupid software

It's not about execution time

[the_real_tigga]

Contrary to what most users say, for me gentoo is great not because of speed (though it's definitely not slower than others), but because of skipping of dependencies.

Linking against only stuff I have (installed) or want (USE flags) is IMHO a great advantage over binary distros, who have to link against everything and the kitchen sink in order to fulfill arbitrary consumer needs.

This may make gentoo binaries faster, quite possible, it definitely gives a nice feeling to see an emerge "actively" NOT linking against alsa, or X11, or libgnome||kdelibs if you specified you will not use it.

There are other reasons why gentoo is great (forums, anyone?), but this is the main reason for me.

Interesting, but I have to wonder.

[Enahs]

I'm not so sure. As they seem to say, you can't count on supposedly identically-configured hardware to be identical.

I'm also not so sure about their claims. Sure, they built the same kernel on the machines, but what kernels did they use on each individual machine? Sure, they used DMA on all three, but did the kernels on all three have support for the IDE controller chipset built in? If not, did all three use the proper module(s)? If not, there's yet another reason this test is completely meaningless. Did all three systems use the same kernel version and were all three patched the same way? Were the same model drives in all three machines? Were partitions laid out in exactly the same way on all three machines? Were all the partitions on all three machines using the same filesystem types? Were those filesystems all formatted in exactly the same way, with exactly the same version of the utilites (unpatched, or using the same patches on each, of course)?

There are too many variables here to make this a valid test. This is about as valuable as a Microsoft benchmark of Windows vs. Linux.

One of the nagging suspicions I have is that they've used performance-tuned kernels (that would more than likely be the case for Mandrake, though it's probably tuned for interactive performance) and more than likely merely followed the installation instructions for Gentoo and installed gentoo-sources. Further, I'd be willing to bet that there are performance tweaks done on both Debian and (especially) Mandrake that were unknown to the reviewers. I haven't worked with Mandrake for a while, but it wouldn't surprise me if they use some of the patches that come with Gentoo's ck-sources. The Planet CCRMA project (a Red Hat-based electronic music-oriented distribution using apt) uses many of these same patches in their kernel, and back when I used Mandrake, they used some pretty cutting-edge stuff.

It's great that they decided to pick on Gentoo; after all, it's always good to take a critical look at wild claims. However, if they're just doing this to disprove some bloke who keeps saying "Gentoo is better, Gentoo is higher-performance" I wish they'd keep it to themselves because they haven't proven anything to me other than that in the hands of an untrained newbie without an eye toward optimizing performance Gentoo gets poorer performance than the average Linux distribution.

For me Gentoo is the only one that really worked

[The Analog Kid]

I switch to Gentoo, once my main machine was up and running again. I use to be a Red Hat user, that was fine up until Bluecurve. Then I was a SuSE user, great except for YaST2, granted its a fine program, but using it to install a system over the internet like that its stupid. It's slow and it takes way too long. I don't know why SuSE just doesn't offer ISOs. Then I was a Mandrake user, Mandrake is a the slowest of all the distros I've tried. I went to Debian. It's like Gentoo with regards to package installation, but I tried to install GAim, I need libspell4 or whatever it was, so I tried to install that, it conflicted with something else, so I uninstalled the conflicting library, guess what GAim depended on them both, and there was no way I was going to sit there and figure it out, it was too much a waste of time. Then came Gentoo. Yes, it wasn't as smooth as installing other distros and I didn't get it right each time. I saw GLIS(Gentoo Linux Install Script), and tried it out, it worked except I had to fix the Grub configuration. Then I decided to do another install on the same machine weeks later. The install went even smoother, and I did not use GLIS. I figured I'd go back to Debian, and give it another shot, because its like Gentoo but without the compile times, didn't like it. Gentoo is my one and only distro. If I need help, there is a forum and there are plenty of people to help me. This isn't a troll or a flame, its just my opinion and my experiance.

The real value of Gentoo

[Anonymous Coward]

i have been running linux for 7 years,
and i learned 100% again what i already know
from running Gentoo
.

gentoo speed

[dukerobinson]

though I have never used gentoo, I have compiled pleanty of packages for linux for my own personal use and I have found that they are usualy a bit speedier that the binary package, however I think a better performance test would have been to use some sort of highly computational software like computing seti packets or something like that, or doing a very complex math problem or something, Opening a spreadsheet is sort of subjective, and is probably not a very good benchmark

I like gentoo

[lanalyst]

I moved to Gentoo after my Redhat 7.3 system borked and wouldn't apply the latest errata via up2date. I moved because I really didnt want to download all those isos and really I outgrew RedHat.. I wanted the lastest versions of applications.. and I probably did shoot myself in the foot installing 'unofficial RPMs' messing up dependencies. Funny, these days I know how to go about fixing such problems.

The one thing I noticed about gentoo was the community.. on irc.. forums.. bugzilla.. the developers are always accessable and anxious to fix or provide help. The user base is top notch and ALWAYS willing to help out a newbie. I think you would be hard pressed to see a 'RTFM' response to a frustrated user's question.

Why is this? Because almost all Gentoo users realize they have learned more about Linux and the culture of Linux with this distro than any other they have tried.

It's not easy coming from RedHat.. it's probably easier coming from slackware but after your system is built - there's a feeling of empowerment and accomplishment. And for all the folks that helped answer your questions, it's pretty easy to jump in and answer help out the next guy..

I like the fact I run the latest versions available. That's the funny part about Gentoo - the Live CD is versioned but after you install, that version becomes meaningless. The interesting part then becomes your favorite desktop environment, version of apache, flavor of kernel, etc etc, etc and isn't that really what's cool about Linux? Try the development kernel.. beta xfree.. development tree of your favorite app. It's okay.. it's in portage.

I believe Gentoo is a wonderful community-based distro that allows folks to learn and grow with Linux. Bugs happen but these get worked through and it usually turns out you dig into an area of the system that you have never worked with before and come away with a few more tidbits.

It was pretty amazing when I first installed.. I had my choice of filesystems, system loggers, cron, and kernels.. mind boggling.

Bottom Line: if I'm hiring for Unix/Linux support and have 2 candidates.. one with a RHCE and another that runs Gentoo at home, who is probably going to have more depth? Who probably has the Microsoft mindset of calling on tech support for everything. Who probably is willing to dig into and understand a problem an offer their own solutions.

Call me a zealot.. piss and moan about compile times.. benchmark apples and oranges.. say it will never make it to the DataCenter.. all that doesn't matter.. Gentoo is fun. Stop by!

(I'm not here to flame anyone's favorite distro.. just pass on my experience).

Re:I like gentoo

[usuri]

I absolutely concur. The level of understanding gained by using Gentoo is a huge benefit....... but having a user base as couteous, understanding and compassionate is the real benefit. Having serious questions (and sometimes silly questions) answered with truth and a smile is much better than a simple RTFM that used to be the standard linux response. (Debrah and Ian..... I am refering to your user base) :)

Moreover, having ultimate control over every aspect of your own system is refreshing. If I want to use openbox, with kicker and login via elogin.... thats my choice.... I could go on and on .... but I'll conclude with a smile and nod to the previous poster..... Herr Lanalyst.

-fomit-frame-pointer

[Pr0xY]

the reason gentoo didnt leave the other's in the dust is because they failed to enable omitting the frame pointer, this is a HUGE optimization as it will free up a register (which is extra important with intel only have 8 regular registers). The result of this is far less read/writing to memory and thus MUCH faster code.

The average desktop user has no need to care about debugging information and it is perfectly reasonable to leavew this out for them.

Try the test again with this on and they'll be more than pleased with the results.

proxy

Re:No. Gentoo is not a performance leader.

[ameoba]

And drinking orange juice makes LSD work better.

Re:right now

[_Knots]

Of course, some realize that distros are good at different things and don't (seriously) insult any of them.

I use Gentoo on my two main boxes, simply because I toyed with it and haven't found a good reason to switch away yet (two just so I have the same environment on my two main machines). That said, my router runs Debian-testing. I maintain Mandrake machines at work, know my way around a RedHat machine... etc. My 486 ran Slack 8 while it was up. My DECStation runs NetBSD (though that's a different issue entirely).

I catch flack for using Gentoo, RedHat, Mandrake... my point is, it doesn't matter at all. Use the right tool for the job - that's what open source, and moreso just diversity in the computing environment, is good for! As long as everything speaks TCP+UDP/IP, who cares? ^^

--Knots;

ravings about debian

[Anonymous Coward]

"Debian would too but the effect will be drowned out by the damage Debian is doing to Debian"

Speaking as a Debian Developer, and Gentoo user (i know, sacrilege and all that).

Debian has a very conflicted nature, it evolves in a mob metality, any changes must first pass the gauntlet of debian-devel-ml where fire and brimstone await.

IF you make it through the gauntlet and have the support of the cabal (sshh, they dont really exist) and ALL the individuals who your changes effect, onlythen do you have a chance of getting the change through.

In the short to medium term it means things a more likely to remain the way they are, which is good if things are already running well.

In the long term, it means debian is doomed, ever hear of the expression "evolve or die" ?

Gentoo's USE settings are the best packaging inovation since apt, however its too radical to make it into debian in the forseeable future.

Debian installer is way better than Gentoo's, but still debian installer only narrowly made it through the gauntlet and much of what they are doing is shrinking debian without violating their precious policy. (they earnt a backdoor key)

Debian is structured in bazzar style, but many bazzar dwellers are trying to modify their stand into a cathedral, these are the elitists who also control the gauntlet.

Debian doesnt have much competition, there arent many non-profit distirbutions at all, i beleive gentoo is listed as non-profit but only exists to make money for drobbins, look at zynot for the explanation.

Whats needed to fix debian is for some real competition to emerge in the same space as debian (and no, mandrake and redhat are not competition to debian). Only true competition will prompt debian to get its head out of its arse.

So....

[Sevn]

We have learned that Gentoo runs like crap if it's
set up by someone clueless. The whole point to
Gentoo is to optimize the crap out of it. And NO,
most users do not run a stage3 install. Almost all
users I know run a stage1, or stage2. That's kinda
the point. But yeah. This test makes perfect sense.
They don't like Gentoo. They had an axe to grind.
Amazing that Debian came out on top.

That IS the point they proved.

[zCyl]

they haven't proven anything to me other than that in the hands of an untrained newbie without an eye toward optimizing performance Gentoo gets poorer performance than the average Linux distribution.

It's probably not fair to call them untrained newbies. From the article it sounded like there were several skilled people participating in the test.

But yes, provided their hardware was identical like they said it was, what they did show is precisely that. In the hands of someone who makes an effort, but may not go to extremes, Gentoo can perform poorer than a binary distribution. If it's true, this is a useful observation, because it can tell some people whether or not Gentoo is right for them.

In other words, it takes more than just compiling the source yourself to get good performance, and if you're not going to do these extra things, don't run Gentoo for performance.

The really disturbing thing...

[Angst Badger]

This is a bit off-topic, but...

The really disturbing thing is that the best time for opening a 32,000 line spreadsheet in Gnumeric was over seven minutes.

This is a perfect example of why, as much as I would like it to be otherwise, I can neither switch entirely to Linux nor recommend others to do so either if they are dependent on any kind of office suite. Excel performs the same task (okay, a bigger task -- my largest spreadsheet is some 59,000 lines) in well under thirty seconds. Open Office runs much more slowly than MS Office on the same hardware and is completely unusable on some hardware that supports MS Office just fine. AbiWord can take thirty minutes just to change the screen scaling with a large document -- MS Word does this almost instantly.

Please don't get me wrong or think I'm a big Microsoft fan. I detest Microsoft, and I've been waiting for Free Software to save me ever since I started using Linux daily in the mid-90's. Common end-user application developers for Linux seem not to use their own products very much or else test them on the latest and greatest hardware. The open source feature set has made great strides since I started using Linux, but the performance of open source office software frankly sucks horribly compared to Microsoft's offerings. This is especially depressing since you'd almost have to try to write software that wastes more CPU cycles and memory than MS software. It doesn't seem to be getting any better, either.

It's not Windows 98 and Office 97 that are too bloated and inefficient for my 1996 IBM Thinkpad with a 120MHz Pentium and 40 megs of RAM -- it's X and Open Office that bring it to its knees. Sure, I could spend a wad of cash to upgrade my hardware so it could handle Open Office, but that's missing the point. OSS is supposed to be more efficient, not less. For myself, this is a much bigger barrier to full adoption than feature parity, which is not a significant issue anymore.

If you're going to address the benefits...

[GMFTatsujin]

The major benefit for me was that Gentoo was the first distro I'd used that gave me the slightest clue about what the operating system was doing, and how the software worked.

I'd tried RedHat, Mandrake, and a few other distros that set everything up for me so I could "just use it." The problem was that in just using it, I had no idea what I was trying to use. I would go looking for software to do x, y, or z, and I'd either find nothing that seemed to do the job, or a jillion different apps that all did the job differently, and I didn't know why to pick one over the other. Add to that the sense of being at the wheel of an out-of-control car every time I wanted to make a change to a .conf file, and my Linux experience was pretty frustrating.

Gentoo was a brilliant introduction into how to install a Linux-based OS. It started me off easy -- here's the command line, here are the commands to install the system, here are the .confs you can tinker with and what they do. It gave me flexability while keeping the results trim. The USE flag is the most amazing option I've ever seen.

Installing Gentoo was more like playing with LEGOs than installing a system, and when I got done with it, I had a computer that I knew, really *knew*. I knew all the init.d services and what they did. I knew what module was controlling what hardware in my kernel *and* how to fix it if it didn't detect properly. I knew all the apps installed, even by their weird names and locations, and I knew what they were there for. I knew it because I built it that way. And I never had to hunt down a dependency or resolve a version conflict. NOT ONCE. Redhat and Mandrake just installed this mysterious Linux Stuff and threw the computer back at me when done. Gentoo got my hands dirty with building it up, but didn't make me jump through hoops to do it.

The benefit was teaching me what my computer was doing when I used it.

*THAT* is how I wanted my computer to run. And it does. Thanks, Gentoo team!

GMFTatsujin

Re:No. Gentoo is not a performance leader.

[maeka]

Actually, the decrease in blood pH caused by the consumption of orange juice increases the secretion rate of LSD through the kidneys and decreases peak serum concentrations.

So, if you like shorter, milder trips, then yea, orange juice makes LSD work better.

Well.....

[Anonymous Coward]

It's difficult to see how a smaller, pentium-4 optimized kernel and customized userland resulting from a custom configuration won't improve the performance of your system. If you were to do a binary-only install (impossible with Gentoo, yes?) that was built with Intel's compiler, perhaps it would work better than one built with the GNU compiler. I'd place my bet on the cutom build though.

Bottlenecks...

[BrokenHalo]

the CPU is not a signifigant bottleneck in modern systems.

I don't know about your machines, but a careful look at gtop on my two home systems (a 2.4 GHz Pentium 4 and a 1 GHz Athlon) seem to indicate the main areas of contention tend to be fairly evenly divided between the motherboard and the CPU. Sure, we can measure CPU and disk drive activity independently of each other, but that usually doesn't tell us a lot. Often logjams can be reduced only by upgrading the motherboard.

Re:No. Gentoo is not a performance leader.

[Squinky86]

I repeat. I agree wholeheartedly with what you said.

...kids today, they can't do drugs right anymore- they have to go and do that stupid stuff that'll mess them up for life.

Relativity

[Kjella]

If you invest a lot of time in learning a distro, you're terrified that it might not be the best, and will spend ridiculous amounts of time insulting the others.

Say that in response to the distro wars, and you'll be at +5, Insightful. Replace distro with OS in a Windows/Linux/*BSD flamefest, I think your score would be -1, Pro-M$ ;)

Kjella

Enough Already...

[khyronmaetel]

Ok, I'm a gentoo user. I'll admit a sizable percentage of our ranks dont know what they're talking about, i'll even admit that most distro "speed" is in the users head. But most of you are missing the point. Many gentoo users (including myself) installed gentoo as an ongoing learning experience. Sure, there's really no difference between the "l337ness" of typing emerge foobar and typing rpm -ivh foobar. But those of us who have taken the time to understand the portage system have learned a great deal. As an aspiring programmer, this was my distro of choice because it enabled me to learn about gcc. Also, i like the idea of (although most install standard packages) being able to beta test bleeding edge applications. While there are a lot of phoney gentoo users who are under the impression that theyre furthering the opensource movement by emerging packages, gentoo's backbone a highly active community of volunteers who are really interested in Open Source. Basically, all i'm trying to say is that any idiot can probably get gentoo installed and working, but the real point is to understand the OS that you've built, and i've found that gentoo helps me and others do this better than package based distros.

Re:Enough Already...

[Anonymous Coward]

Ok, I'm a ex-Gentoo user.

According to you Gentoo is a distribution for beginners. ;)

I've started using Linux using RedHat 6.2. I know. RedHat sucks. Every time I try to switch to another distribution I'm hit by a deja vu feeling: "Man. That bug was already fixed X years ago by RedHat."

RedHat 9.0 REALLY SUCKS. So, I switched to Gentoo. After spending A LOT OF TIME installing it from stage1 I started looking at /etc scripts and files to configure Gentoo according to my needs, only to discover how crude are the Gentoo configuration system. And it still has the same bugs that were already fixed by RedHat X years ago.

My machine was freezing every day. How I fixed it? I started using the vanilla kernel instead of the Gentoo patched kernel.

I'm not the only one who got disappointed with Gentoo (Reasons for forking a Linux Distribution).

Last night I switched my machine to Debian.

Re:Enough Already...

[Al-Hala]

I'd mod you up if I had available points.

I've played with the live CD's of most companies, and that was the extent of my Linux experience, until I jumped in and installed a Stage 3, then Stage 1 Gentoo box.

Despite the STEEP learning curve in my particular case (ATI hardware, infamous on board sound, and other issues) I think it was the right choice to get the understanding I obviously didn't get with the Live CD's.

Re:No. Gentoo is not a performance leader.

[dh003i]

Omit-frame-pointer is not a regular optimization.

Irrelevant of such, its one that many Gentoo uses need. Many Gentoo users don't need stack traces to hand to a developer, period. You also ignore a common optimization used in Gentoo -- -Os, which optimizes for size, producing smaller binaries which load quicker, and take up less RAM.

compiling for different architectures generally makes very little difference on any platform other than compiling for i586 on a Pentium

This is an unqualified statement that doesn't hold. It might be true for some very generic functions, like word-processing and internet browing. However, for many scientifically intense functions, like phylogenetics, architecture-specific optimizations do matter.

I will agree with you that the specific optimizations don't make a huge difference in most cases; in some, however, they will. And of course, the other benefit is that software is only compiled with support for stuff you want, and that you have absolute control over what's there and what's not from the start.

Re:No. Gentoo is not a performance leader.

[Anonymous Coward]

Sheesh, another Gentoo fanboy. You all live under rocks around here or something?

BTW, I don't think these guys are into phylogenetics. Can't imagine why - I'm sure it's what everybody uses their Linux boxes for.</SARCASM>

Re:No. Gentoo is not a performance leader.

[Anonymous Coward]

Yeah. Sure.

Can I call you "Mr. L337 Haxx0r"? Please?

Re:No. Gentoo is not a performance leader.

[0x0d0a]

Look, Squinky86. I'm not simply pulling this out of my ass. I've had to cover this back in university. I'm not a gcc developer, but I have sat down and gone through generated assembly from the compiler, and have spent many hours tweaking software in every way possible to make it run at a reasonable rate on my old P2/266. There are a very few pieces of software for which arch flags make a measurable difference (as a later poster noted, gzip is one). As for individually specifying flags, I'd be facinated to know what you're trying to use above -O3. You can use -ffast-math. It's unlikely to provide particularly useful results. Approaches like this have been done for a long time (see libmotosh on the PowerPC) -- they can cause the rare, PITA to find problem, and any libs or programs that really need the speed increase have probably done custom work that's even faster than anything you're going to pull with ffast-math (libfftw, for instance). You *might* get some performance gains on povray...but most folks I know compile povray themselves anyway, since it isn't packaged by, say, Red Hat. -fstrict-aliasing is a *very* balsy flag to use if you haven't actually written the software yourself. It's a pretty safe bet that building a lot of unknown software with -fstrict-aliasing will break it. There's a good reason strict aliasing is off by default -- valid C programs (easy ones to write, too) will die with this option, occasionally and in odd ways. You're a damned fool if you use this on *any code* that you did not write or explicitly says that it was written to allow this optimization. I just finished talking to an optimizing compiler designer Thursday who reinforced my feelings about aliasing-dependant optimizations -- they're almost always a bad idea, since the small speedup isn't worth the random problems that you can very very easily induce. -fomit-frame-pointer can produce a small benefit, but surprisingly small, and makes tracking down any crashing bugs or requesting help with a crashing bug infeasible.

If you know how to and properly configure your compiling flags, the speed gains are tremendous

Bullshit. The vast majority of software I've run benchmarks on will never see less than a 10% performance gain (and that's being *very* generous...most will see no measurable change) with anything other than the default -O2 or -O3.

Oh, hell. I was a lot like you not so many years ago, sure that I could speed things up if I just found the right ways to manipulate the code. The only cure for it is actually sitting down and benchmarking things yourself, since you're sure that everyone else is doing something wrong.

Go ahead, you'll see what I mean. Try building libs that tie up a lot of CPU cycles in multiple apps like libjpeg -- that's where you're going to see your best payback for any optimizations. Time a couple runs.

but if you just try gentoo, the learning experience and speed gains are very noticeable.

I think I've adddressed speed gains. As for learning experience, Gentoo is not synonymous with compiling software from source (from that standpoint, Slackware and similar distros blow Gentoo away). I've never bought into the "learning experience" claims -- let folks start out on the GUIs their distro maker provides, and then, regardless of distro, you can quite happily find out what's going on.

This is not to say that I don't think Gentoo is a worthy distro. I'm a bit of a package management aficiado, and emerge certainly interests me. However, the kind of sweeping claims I see the occasional Gentoo user make on Slashdot are ridiculous. The general-purpose Linux distros are all fairly close together. Distro fans tend to be produced when someone fails to understand how to properly use a different distro (or got accustomed to one), or has sucked down some false claims from other folks, or just don't want to consider that the distro they've sunk lots of time into learning isn't far better than any other choice.

Now, if you happen to like Gentoo, go for it. But like it for the legitimate reasons, not inflated false ones. Don't make exaggerated claims WRT to it, because misinformation certainly doesn't help out Linux folks in the long run.

Re:No. Gentoo is not a performance leader.

[Sir-Tez]

Say, Here's MY CFLAGS list:

CFLAGS="-march=athlon-xp -m3dnow -msse -mfpmath=sse -mmmx -O3 -pipe -fforce-addr -fomit-frame-pointer -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -falign-functions=4 -maccumulate-outgoing-args -ffast-math -fprefetch-loop-arrays"

The fact that they only used a mere one optimization doesn't show you just how fast Gentoo can go. They need a hardcore gentoo optimizer in there for the gentoo box, someone that knows what they're doing. Gentoo is completely customizable but the customization is dependent on the person doing the installing. Hence, I say the person doing the gentoo install is NOT informed on optimizing gentoo and thus renders this test invalid.

Try it out

[0x0d0a]

[shrug] I'm open to new ideas. Convince us. Compile a couple pieces of software with and without your favorite extra flags versus just plain old -O3 and post your times. My experimentation in the past hasn't turned up any significant improvements, but you're welcome to try on more pieces of software and see if you can find major improvements. Avoid gzip, as this is a well-known exception that many people *do* custom-build with an arch setting regardless of distro.

Normally, I'd agree with the "support for whatever you want" argument, but I've spent phenomonal amounts of time in the past trying exactly that -- shaving down binaries, etc. It just plain doesn't make a difference. My favorite example is rxvt. Rxvt is a beautiful, svelte piece of software, and the fastest xterm clone out there. You can comple out almost everything in the package. But, you know what? Linux's loader and VM system are pretty smart -- they only store one copy of the binary in memory. The speed savings you get are pretty small. Most things these days have runtime-settable featuresets. It just really doesn't make sense to try to poke and prod at your system for the extra 1% of speed, because it's just a waste of your time. If you want to make a much more significant difference, try rewriting a chunk of mozilla or something to use a more intelligent algorithm in one place or another. You'll see much more appreciable improvements. I've rewritten core code in gtk-gnutella and dillo for significant performance improvements -- far, far greater than I'd see recompiling software with different options, actually taking less time, and benefitting all Linux users instead of just myself.

Why I don't like up2date

[0x0d0a]

Well, I haven't poked at it recently, since yum really wowed me.

I've seen it crash an uncomfortable number of times. Last I remember, I had to register the darn thing, giving Red Hat a sellable (forced confirmation) email address. The GUI on the thing (admittedly, not the only way to use the thing, but its CLI is much less comprehensive than apt or yum) is terribly unresponsive -- it blocks when there's an operation going on in the background. Up2date requires a custom server for doing what yum and apt can do with regular ftp and http servers. There are many third-party repositories of software for RH which are only supported with apt and yum (come to think of it, I'm not entirely sure that up2date can even handle multiple sources).

up2date runs slowly (yum is also written in python, but it's much peppier). Yum has pretty intelligent fallback downloading. up2date requires an additional flag to run in console mode. up2date can't display package headers.

IIRC, up2date doesn't cache, as apt and yum did, and just doesn't provide all the functions that apt and yum do.

I've always felt that up2date is RH's biggest out-of-box weakness. There's a good reason that the third party RH providers (freshrpms, fedora linux, newrpms, planet ccrma, atrpms, jpackage, macromedia) all use a different distribution system than up2date.

Gentoo vs the rest of the world.

[Anonymous Coward]

Forget it, I actually wrote something up here mildly intelligent, but its not worth posting, it will be lost in the depths of intelligence of the armchair experts.

The biggest advantage is...

[Anonymous Coward]

on platforms where gcc is really, really bad, e.g. alpha, itanium (ia64) and MIPS. Unfortunately there's no gentoo port for my itanium system, but if there was I'd jump to it, because there are so many programs I use every day (image magick, ghostview/gv, and good old bzip2) which run a LOT faster (we're talking factor of 2, sometimes more here) if I compile them myself using the Intel C/C++ compiler.

Right now, I basically have a debian unstable system which I feel free to clobber in the name of performance, and it works OK. Using gentoo, assuming it would one day support itanium and the Intel C/C++ compiler, would be a much more elegant way of doing things, and get me even better performance with fewer hassles.

Learning

[xRelisH]

I must say, that Gentoo is a great Distro for actually learning some of the things of the Linux environment.
Gentoo was/is my second distro (I tried redhat first), I was intrigued by the idea of doing everything from almost scratch.
Starting from Stage 1, I was able to learn quite a few things that I probably would not have learnt till some time if I had not decided to do a Gentoo Stage 1 setup as you're somewhat forced to learn the basics and have a fairly good foundation to go learn more stuff yourself.
Although I am sure there are other distros out there (including linux from scratch, if that's even technically a distro ) that are like Gentoo, but I find it to be a rather good Distro.
However though, I may decide some time from now to try the "from scratch" route, and perhaps learn even more. Although, I think the think that will keep me with Gentoo, like others have said is Portage, I think it would be great if there was a well maintained Portage-like system for any distro, as a piece of software, even though it's probably possible to get portage working on other distros, but probably not without some major tinkering.

Re:Learning

[marcushnk]

I'll second that line of thought.. and up the ante by saying that Gentoo rocks as a gaming platform as well.. It works fuggin well for me.. and I'm a linux n00b :-)

Let's take a look at your claims

[0x0d0a]

They need a hardcore gentoo optimizer in there for the gentoo box, someone that knows what they're doing....Hence, I say the person doing the gentoo install is NOT informed on optimizing gentoo and thus renders this test invalid.

[sigh] Nobody ever listens to me [Dark City].

Okay, let's take a look at how informed you are. First of all, -mathlon-xp implies -m3dnow, -msse, and -mmmx. -O2 or -O3 is default already on most systems. The only differences -O3 produces is -frename-registers (which does essentially jack on the x86 line) and inlining (which tends to produce very minimal or negative benefits, given the fact that cache misses (which this aggravates) are far more of a timesink for most programs than setting up and returning from function calls). -pipe produces no runtime benefit, though I leave it in my own flags. -fforce-addr, -frerun-cs-after-loop, and -frerun-loop-opt are implied by -O2 or -O3 already. -falign-functions=4 is considered a slowdown for the Athlon line by the gcc team relative to the default (64 on current gcc). I haven't tested -maccumlate-outgoing-args, and I'm not familiar with what it internally does -- the only benchmark I could google for indicated a slowdown caused by it. -ffast-math is a decision of dubious value. Very little code uses floating point math, so ffast-math rarely has an effect. The code that does and actually cares about this degree of performance generally has native implementations that are faster than -ffast-math, since they're special-cased. This can cause software breakage. (We already saw the realization that these sorts of optimizations are of dubious value with Motorola's LibMotoSh for the PPC). -fprefetch-loop-arrays is implied by -O2.

The overwhelming majority of code does *not* have an #ifdef __SSE__ with alternate code.

Basically, the only seriously useful flag you used is that which all distros use -- -O3 (and I generally feel that -O2 is a better choice on modern processors, where cache is so critical). -march=athlon-xp can help, but it's unlikely to make a measurable difference on any but a very few pieces of software. Most distro vendors already benchmark and ship versions of software that benefits with a different arch -- look at RH's different RPMs. -fomit-frame-pointer is arguable -- but you're going to probably see *well* under a 10% performance difference, and you have no ability to track down crashing bugs or send in useful bug reports. -ffast-math can cause breakage, and provides little benefit for almost any package (one exception is povray -- it's a floating point heavy package that tries to be portable). I custom-build povray, but then RH doesn't package povray anyway, so that's not too much of a concern.

Anyway, my point is not to criticize you. I spent my days with a three line CFLAGS string as well, sure that I was producing nicer and better code than anyone else. Then came my benchmarking days and a compiler class and some days picking apart gcc-generated assembly...and I realized that I really wasn't gaining anything.

If you like Gentoo, do it for the legitimate features that it provides (like emerge), and not for some fanciful performance improvements.

That was a very poor performance test..

[cbreaker]

I'd have to say that this was one of the most poor performance comparisons I've ever seen.

Not only did they simply seem to download and install gentoo without even thinking about configuring it properly, they did a few little tests with very little explination about each test.

While I'm sure the scores are accurate for their tests, they obviously did not actually use Gentoo the way most people do: Optimized for their particular systems and their software.

Where Mandrake version whatever will always perform exactly the same and function exactly the same, Gentoo grants the ability to compile and configure software exactly how you want it.

Not to mention that the reason I choose Gentoo isn't because of performance - it's because of the extremely simple and easy interface to keeping your system updated and installing new software. Portage is the name.

Double byte character input!

[ahfoo]

It looks like I'm the only one to bring this up, but as a native English speaker that frequently has to work in Chinese what caught my eye about Gentoo was its alleged ability to made the bleeding edge character input systems work.
I haven't actually tried it yet, but I'm very intrigued with their on-line instructions showing a working copy of Chinput which is very close to the kind of Chinese input support we see on the Windows desktop and unfortunately this is one area where Linux desktops can really lag. I've had success getting Xcin to work on various distros including Knoppix, but the NCurses look is really not what end users expect in a key-in system these days.

Well I must say

[gunix]

that I couldn't care less about wheter it is a source or binary distribution.
What counts is to have the latest (and greatest) version of the programs. They tend to be better more bugfixes etc. (and new ones of course)

I run debian stable right now, but as soon as I get a new harddisc, I'll turn to debian testing.

With a source dist I guess you are even more on the bleeding edge, right? Or at least, could be...

Speed

[teklob]

Sure gentoo can take a while on older hardware, but on new hardware, such as an Athlon XP 2400+, I can build X and Fluxbox in 84 minutes AND have whatever support I need compiled into it and not what I dont. Also portage is super easy.

oh FreeBSD, yeah it's great

[DrSkwid]

FreeBSD Passes 9000 Ports

Re:No. Gentoo is not a performance leader.

[Anonymous Coward]

You're arguing with a Gentoy fanboy, and you're using actual facts and figures to back up your arguments. You may as well give up now and save your sanity. These Gentoy users wouldn't understand a real world fact if you beat them with it (Lord knows I want to)

You missed the bus

[Anonymous Coward]

Maybe you should go back and read the Gcc manuals and figure out what all those fiddly little flags do, because it is clear at the moment that you have no idea.

If you can't work out why -mcpu=pentium4 -march=pentium4 -mmmx -msse2 is redundent, or why -ffast-math is a bad idea (Or at best, a waste of time), then you have no business compiling your own software.

I hope to God you're not actually allowed near anyones production server.

prelink

[Anonymous Coward]

What about prelink? I believe that RH already comes with it by default, what about Debian and Mandrake?

Gentoo doesn't, which would make the test _VERY_ unfair.

Re:prelink

[lanalyst]

prelinking in gentoo

Re:prelink

[Anonymous Coward]

I know that link, and I use prelink, what I was impling was that prelink doesn't come by DEFAULT. GET IT?

Valid tests, not likely; Gentoo's real calling.

[mindmaster064]

While it really isn't likely that Gentoo is more than 10% faster than other systems in any given operation there are reasons I don't feel these numbers reflect anything.

First, the tests fail to mention kernel versions they're running or the filesystem of the machines (which has a big impact on processing efficiency, different fs being better or worse at different things) and they didn't even mention what version of X11 or gnome they're even running on these boxes. Are we testing Gentoo? Gentoo is portage, gentoo-sources (and the other modified kernels), ebuilds, prelink, evms, XFS, and a host of other features the others simply don't have.

Second, no comparisions are made on i/o bound tasks (reading/writing/copying files). Things that do not stretch resources in linux may simply be preempted out of importance. (Linux puts higher priority on a soon-to-be-overstuffed disk buffer than the display/computation of a spreadsheets). It would be hard for ANY of these distros to even compete on i/o bound tasks mostly because of Gentoo's EVMS or XFS options. EVMS is simply the best Software RAID I have ever used. XFS seems to be one of the filesystems of the future as well due to it's i/o guarantees, journalling, and sickly-disgustingly-large filesizes.(I feel the jobs XFS will do simply cannot be done by anything else at this point).

Thirdly, another advantage is simply not having 3-4 disks worth of iso install that you will NEVER need. Mandrake and Redhat (though not tested here) are completely bloatware. They don't install what you need and work properly, they install EVERYTHING or work poorly (because the distro makers hide some silly dependencies). Gentoo's "emerge -p packagename" is a whole hell of a lot friendlier. But, if you don't use the deep options it's possible to lose a lot of the "roll your own" advantages. Gentoo is subject to a high degree of user stupidity, it is possible to negate all of the advantages of source-based distro simply by being careless.

Lastly, Gentoo's major advantage -- it's best.. the Gentoo users! That's right. Gentoo's forums are so good (and full of know-how types) that put virtually any goofy problem you can think of is within searching distance. I simply have found Gentoo's support forums probably the best repository of linux information on the net. I don't say that lightly either. Something to be said about being able to get problems resolved quickly and correctly -- and having it at your fingertips. I like to go to one place for answers, and likely you do to.

As you can see, it's very hard to get a good feel for what Gentoo is by reading that review. I would encourage any techie freak to take her for a test drive. It's not for everyone and not even for most, but it's still a very different experience in comparison to other linux distributions. (I have used Redhat, Mandrake, and switched to Gentoo -- I have not had any need or thought of moving to something else; my own 2 cents).



-Mind

And debian is the most free ! Go GNU !

[Anonymous Coward]

;) Gnu/Linux/Debian all the way !

Default CFLAGS In gentoo is not common

[fire-eyes]

In a "typical" install of Gentoo, the CFLAGS are changed significantly.

Mine are quite aggressive, and possibly redundant and maybe silly. But it is definately faster than any other distro I have ever had on here, in every single way.

Please, spare the comments about what may be redundant etc. Works For Me. BRILLIANTLY!

CFLAGS="-march=athlon-xp -O3 -pipe -fexpensive-optimizations -fomit-frame-pointer
-D__SMP__ -mmmx -msse -m3dnow -mfpmath=sse,387"

Gentoo is FUN!

[HuskyDog]

I use Gentoo because getting it working properly is FUN! Back when I started using Linux (1994?) getting it to do anything at all was a fun challenge. These days I think there is a good chance that my mother could install Mandrake. Where's the fun in that?

I used to spend hours fiddling around with the printer config files trying to find the best ghostscript command options, never mind trying to get X working properly. I run Mandrake on systems where folks will complain if it breaks, but on other machines I have hours of amusement tweaking Gentoo. Recently I tried to install it on a 66 MHz 486. It took 6 weeks of almost solid compiling before I could get KDE working on it! Now I'm looking for a 386 and I have a SPARC and Alpha which I am going to try.

I haven't had so much fun with Linux for years!

ever take stats course?

[Anonymous Coward]

you measured a semirandom quantity and you only measured it once? you are supposed to measure it like 10 times and take the average

Re:No. Gentoo is not a performance leader.

[iangoldby]

-fstrict-aliasing is a *very* balsy flag to use if you haven't actually written the software yourself. It's a pretty safe bet that building a lot of unknown software with -fstrict-aliasing will break it. There's a good reason strict aliasing is off by default -- valid C programs (easy ones to write, too) will die with this option, occasionally and in odd ways.

Getting slightly off-topic, but if you care about strict aliasing optimisations, you can use Fortran. The language is structured in such a way that the compiler knows exactly what aliasing is going on, and can optimise accordingly.

You've never used URPMI, apt-get, up2date.

[leonbrooks]

Even with the compiling, it takes less time to install some stuff (eg nmap) than it would take to locate the relevant .rpm.

This is on a Celeron 300A with 64MB of RAM, and had to update the menu entries for 13 (!) WM's (dd wanted to try them all):

[root@aiyana root]# time urpmi nmap-frontend
To satisfy dependencies, the following packages are going to be installed (1 MB):
nmap-3.00-2mdk.i586
nmap-frontend-3.00-2mdk .i586
Is this OK? (Y/n) y
installing //home/leonb/1/Mandrake/RPMS/nmap-3.00-2mdk.i586.r pm //home/leonb/1/Mandrake/RPMS/nmap-frontend-3.00-2m dk.i586.rpm

Preparing... [row of # deleted to make /. junk filter happy]
1:nmap [ditto]
2:nmap-frontend [ditto]
24.75user 21.40system 1:17.24elapsed 59%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (2583major+11904minor)pagefaults 0swaps

So hard, ooh, all that typing - tell you a little secret about optimisation, too: optimising to Pentium level (Mandrake's default) gets you 98% of the speed gains you'll ever get from the technique, except for a very few special maths-intensive programs. Some of the Cooker denizens spent a bit of time building distros optimised for their Athlon, P4, etc and with the exception of a handful of apps like X, glibc, the kernel itself and codecs the difference was squat (in fact, some apps ran slower).

(edits /etc/lilo.conf to boot urpmi)

[leonbrooks]

I win. (-:

(edits urpmi into the BIOS)

I win again. (-:

Gentoo is still my favorite :)

[Anonymous Coward]

I'm new to linux and like Gentoo very much. The performance test doesn't proof anything, nada, zilch. Gentoo has a nice flashy website and a great non-elitist community that is young at hart. Portage is exactly what I was looking for to keep my system up-to-date with the latest and the greatest. I don't know much about optimization, but I just don't care what the big boys say about Gentoo. Gentoo cannot be compared with Debian, Redhat or any other distro. Gentoo is simply unique.

Gentoo is simply unique

[Anonymous Coward]

Gentoo has the most flashy website. All the boring distros like Debian, Mandrake or Redhat can't claim that. The Gentoo forum really kicks ass. Smell you later!

I switched BACK

[Captain Rotundo]

I tried gentoo. I switched back to Debian within two months. None of the claimed benifits were there and it was way less user friendly, emerge would fail with cryptic messages that required investigation, and this wasn't even with 'unstable' ebuilds. - I've been running a debian box since ham, and have never had apt fail even on unstable (of course except for cases where I forced things it didn't want :)

Re:I switched BACK

[LinuxGeek]

The only real problem I have had with portage (emerge) was caused by a prematurely released tcl that pooched the requirements for tclx. The discussion took place here. An overheating laptop was crashing during long compiles too, but not the fault of portage or gentoo. Overall my impression is much better than I had trying to track down rpm dependancies with redhat and mandrake over the years. I have run most of the popular distributions since 1994 and still have most of my servers on redhat, but like gentoo for desktop use and am now starting to use gentoo for specialized server installs. It is much easier to run portage/emerge remotely than either up2date or rpmdrake. Apt is pretty good, but I never caught the debian bug like some of my friends.

I'm not knocking debian, just saying that portage has been great for me so far. You probably had an easily corrected problem with gentoo but are more comfortable with debian so why learn more methods of doing the same things.

Optional dependencies are the real benefit

[CheshireCat]

I used to use debian, until I tried out Gentoo. I find that the real benefit of a source-based distribution is packages with optional dependencies. Gentoo can build emacs with optional Gnome support, or, if I choose, with no X support at all.

Sometimes this can be achieved in binary distributions by having multiple version of a package, but it's hard to reach the level of flexibility that Gentoo is capable of. Suppose a program has 4 optional dependencies in Gentoo, and they're independent of each other (any possible combination of them can be used). To offer the same choice with binaries, you need 16 packages.

Excuse me, but...

[Anonymous Coward]

...(from that standpoint, Slackware and similar distros blow Gentoo away).

Commander McBragg "There _is_ nothing similar to Slackware, young man." /Commander McBragg

Re:No. Gentoo is not a performance leader.

[0x0d0a]

Heh...that should be "never see more than a 10%" not "never see less than a 10%". :-)

Re:No. Gentoo is not a performance leader.

[Squinky86]

Hey, I enjoyed your post and agree with it. Most compile flags do nothing; the main things that matter are the cpu speed and disk i/o. I've said that before.
Also, I don't recomend Gentoo for most users- I am not a fanatic, I just like configuring things myself so when something messes up, there's only one person to blame (go slackware!).
What you should note is that there ARE performance gains, though slight. I use gentoo because I like editing the code then building it specifically for my system. I didn't know there were this many gentoo-haters out there, but I definitely see only good things from this distribution. No need to bash it nor me. That's getting a little childish.
What I was saying was Gentoo is definitely a distribution to try for new and intermediate users to get to know how their system works and what can be done (e.x.- what services are unneeded) to speed it up. Gentoo is a great learning distribution and very easy to keep up to date- and that is a fact (hey, all you people are talking about now is "This Gentoy can here all the facts you can throw him and not move."). Other distributions are just as good, I just don't see the point in flaming a distribution that is equivalent to the others that's easier for advanced users to edit.

True -- up2date does suck

[0x0d0a]

This is why most serious RH users (anyone who sanely wants to use freshrpms, fedora, etc) use apt or yum. up2date is really atrocious. But while RH doesn't package apt/yum, these third party folks *do* do so, for each new RH release.

Re:No. Gentoo is not a performance leader.

[0x0d0a]

Also, I don't recomend Gentoo for most users- I am not a fanatic, I just like configuring things myself so when something messes up, there's only one person to blame (go slackware!).

I'm all for folks recommending things they enjoy. The only thing that upsets me is the repeated number of Gentoo users that come on Slashdot or Usenet and make claims that simply are not true about the benefits distro. I see a *ton* of Debian fans pushing their own distro, but usually it comes down to someone claiming an ideological difference or liking the distro's packaging policy -- you don't see a bunch of claims that Debian is much faster or more stable than other distros. It hurts all Linux folks to have a lot of false information floating around, and it really doesn't do Gentoo any more good than it does to have Linux folks making outrageous claims about Linux's capabilities.

What you should note is that there ARE performance gains, though slight.

[shrug] That's probably true. -fomit-frame-pointer, at the least is pretty sure not to hurt you aside from the inability to debug.

I didn't know there were this many gentoo-haters out there, but I definitely see only good things from this distribution. No need to bash it nor me. That's getting a little childish.

I don't think I was bashing Gentoo -- I specifically said that it was a "worthy distro". I may have been slightly harsh towards you, simply because of my frusteration with the behavior of a large number of Gentoo users on various forums, but I don't think it was particularly out of line, and I pointed out that I've been in the same boat WRT compile flags.

What I was saying was Gentoo is definitely a distribution to try for new and intermediate users to get to know how their system works and what can be done (e.x.- what services are unneeded) to speed it up.

[shrug] I'm not saying it's *bad* for them, though every distribution I know of (well, I don't know precisely how Debian deals with source, so perhaps not Debian) provides a pretty good learning system. On rpm based systems, you can just create ~/.rpmrc, modify buildarchtranslate, fill in an OPTFLAGS field, and then use rpm --rebuild on a SRPM to do the same rebuilds that you'd do on Gentoo.

My other real irritation is the number of (a long time ago, mostly Slackware, now mostly Gentoo) users that run around telling would-be Linux gurus that any advanced users *clearly* use a system in which they build their own packages. This tends to put these people in a position of feeling that they need to use a particular distro to prove their tech balls. It's not a particularly friendly environment for any new users. (They ignore, of course, the fact that Linus uses KDE on SuSE, and Alan Cox GNOME on Red Hat -- and both of them seem comfortable).

Does that make Gentoo bad at all? No. Of course not. I'd just like to keep the twin nastinesses of "Bob sucks, because he only uses distro X" and "Distro Y beats the shit out of other distros when it comes to performance/stability/other inflated claim" to a minimum.

Re:No. Gentoo is not a performance leader.

[Squinky86]

Wow, I agree completely with all of that! And Gentoo users are quite active in spreading the word about how neat and customizeable Gentoo is. I agree, it is getting kinda out of hand. As for "not recommending Gentoo to most users," most people around here don't use Linux, and I start them out on Mandrake. To tell you the truth, I'm happy with any up to date distro and some light wm so I can get my programs and debugging done. Gentoo's customizeable and I can set it to do whatever I want, which is why I've recently started using (was redhat $ mandrake user). It's easier for me. Thanks for you comments- I find them truthfull and informative!

nonsense benchmark

[Jacek Poplawski]

The first lesson of this test is that minor version numbers can apparently make a large difference. If you're using Gnumeric 1.0.13 with large sheets and waiting a while for them to open, it's probably worth investigating version 1.0.12.

Ooops. Two different versions on different distros. Let's fix it. Oh, still difference. Maybe it's because different GNOME-libs versions on different distros? OK, let's fix it. Oops. Different GTK is used. Oops. Is same version of glibc really matters? What is that glibc anyway... Oops, first distro was tested with KDE running, maybe we should switch to IceWM there too? Ooops, do you think running Apache may slowdown gnumeric?

Seriously, this is just nonsense. If you want to compare something, first you need to understand how it works.

Re:nonsense benchmark

[Brane2]

And even more, why didn't he publish the CFLAGS and CxxFLAGS used ?
What is a point of using -O2 if one misses the rest ?
I have discovered the Gentoo twoo weeks ago, after trying with just about every Suse and many other distros.
I used "-O3 -pentium3 -msse -mmmx" and I don't need a chronometer to know that this thing runs circles around Suse 8.2 in KDE...

For me Gentoo's not about performance.

[mystran]

Gentoo's not really faster on my hardware. That's identical hardware. Some things like mplayer might ARE faster than binary packages, but most of things are just the same. Won't notice the difference.

Now, that's not the reason I run Gentoo. The single reason I use it is Portage. It's just as easy to use as apt-get in Debian with 2 advantages.

First, there's USE-flags. You get to choose what you want. I don't want MySQL or KDE support in any packages, so I don't have them. If something supports PostgreSQL, that's compiled in instead.

Second, it's somewhat more up-to-date with newer versions that Debian. For servers I'd still use Debian, but on my desktop I like to stay current. Not bleeding edge, but current. Also there's less trouble with Debians policies about Licenses.

The package management is the best I've seen so far, so that's why I run Gentoo. My previous distro was LFS, but Gentoo gives me all the customization I want, with package management.

myslef

[Anonymous Coward]

i use a hand-modified slackware 9 system. i would like a totally hand-rolled system but that seems to be out of my abilities / ambitions at this time. i prefer to install programs from scratch because i prefer to have the source code so i can see how things are done (i mainly use my box for programming).

however, convenience is nice. i am debating whether to switch back to debian (i flip about every 6-12 months between slack & debian) when i go to university in a couple months so i can take advantage of fun tricks with apt.

well, whatever floats your boat.

Re:Gentoo loses! (-O2 vs -O3)

[LinuxGeek]

The P4 based celerons have 128K of l2 cache just like the original celeron 300a. Only the FCPGA2 ( last of the P-III) based celerons had 256K of cache.

Optimizing for the northwood P4 which has 512K of L2 cache and then running the resulting code on a 128K L2 celeron does impact performance a lot.

Really?

[SiMac]

Even with the compiling, it takes less time to install some stuff (eg nmap) than it would take to locate the relevant .rpm.

Not with something like Debian's apt-get (fine, it's not RPM), yum, or Red Hat's up2date.

YES optimize for size!!

[Anonymous Coward]

I've found that optimizing with -Os improves the performance of Mozilla by an order of magnitude on my Celeron (P2 core). Of course, I have to edit the ebuild because the morons strip all settings and supply a default of -O2. Drives me nuts.

IDE?

[Anonymous Coward]

Apparently someone isn't aware of how much CPU time using his IDE drives is eating....

I use Gentoo because:

[iplayfast]

1. I can make things work in it. (sane with my old scanner just doesn't work with mandrake, but Gentoo works just fine. I only had to follow directions).

2. Portage, makes installing software a snap.

3. I can play with the optimizations. It feels faster to me.

4. I'm usually using the latest software (2.6-testing2 kernel to write this).