Benchmarks Of The Gentoo-Based Calculate Linux

[mirv]

I think it's pretty funny to see Gentoo fans arguing one of its strengths is saving disk space and memory, while at the same time apparently a fast quadcore is needed to fully enjoy it.

Gentoo might be neat if you like to fiddle around for the sake of just fiddling around, instead of getting work done. The ricer comparison isn't that far off.

Well now that's trolling, and completely off base.
Who said you needed a quadcore to enjoy it? It might help make compile times less, but that's all. I run gentoo on an atom for a htpc. Works fine.
Go back to your playpen.

[pingufunkybeat]

Gentoo is neat if you like to fiddle around AND get work done.

Some elitists seem to think that nothing except Ubuntu or Fedora could ever possibly work, but there are actually many distributions out there, and you can get work done on most of them.

Assuming that a distribution is not good for work just because it's not called Ubuntu or Fedora is more elitist than any idiot on funroll-loops.

[deanjo]

Stop the deamon, yes, but all your apps still link against it and you can't remove it. In general, you have lots of libraries lying around that are not needed.

Like people have pointed out, it doesn't usually hurt if you have enough memory but it is a bit annoying in some cases. You shouldn't HAVE TO install GStreamer if you don't intend to use it, for example. And GStreamer should not have to pull in gconf, which doesn't need to pull in all of gnome-base, which pulls in other stuff, and all of this gets loaded every time you start any X program.

It's not a huge deal, true, but I don't need it.

Memory isn't a concern in reality. The mandatory packages that get installed even if you choose not to install pulse in openSUSE for example total less then 1 MB (612k libpulse0) of disk space and ram wise are not even present no daemon loaded at all. (even then you can mark the package as taboo and ignore any dependency warnings as your not going to be using pulse in the first place).

[monraaf]

That's a bit disappointing. What where your CFLAGS?

[yotambien]

Nice, I take a nap and I'm welcomed by a wall of posts...

I want to note that I can install Compiz withouth installing Gnome. Or install mplayer-mt instead of the regular version. PulseAudio as such is not present in my system, except for three libraries that contain 'libpulse' in their name, accounting for about 600K. Sure, there are choices I can't easily make, but I'm yet to see an example that really matters and would justify having to compile everything--as opposed to compile just the offending project whose binary packages don't suit your tastes.

In my case I only compiled in the past, more or less regularly, the kernel, the radeon drivers and a hand of software unavailable otherwise at the time. None of this was for fun; as soon as the distribution packages were offered or updated enough I happily switched back to them. Compiling stuff requires an effort and it's not a reasonable way to distribute software to a large audience. Gentoo users can make up reasons to rationalise their choice, but so far I haven't heard anything solid except the wish to tinker. Which is perfectly fine, and honest for a change.

[pingufunkybeat]

Which distro are you running, yotambien?

I would add the excellent documentation and community to Gentoo's strengths. And the compilation itself is automated -- that's the whole point.

[deanjo]

That's a bit disappointing. What where your CFLAGS?

Whatever I dictate the defaults to be when setting up my environment in .bashrc or specify them when building them on the build service.

[locovaca]

This is why Gentoo makes a great learning distro. It exposes the Linux internals and lets you learn about it. It will assist you, but it doesn't have shiny wizards who do magic for you, you get to see how things operate behind the scenes.

After a while, some people switch to another distro with more knowledge under their belts. Others like the Gentoo way and stick with it.

This is what I did, more or less. I stuck with Gentoo for 4 years on the desktop and 5 on server. What did me in was the circular dependencies of use flags on some packages and the breakage of my system (liberror infuriated me to no end). After I spent 3 hours in DDD one Saturday morning trying to figure out why my USB drive wouldn't automount in Gnome I decided that, instead of fighting my system for 3 hours every night when I got home that I'd rather do something productive after fighting my work computer for 10 hours. So away I went to Ubuntu. The knowledge I got from Gentoo certainly helped.

My systems under Ubuntu are no less customized than they are under Gentoo. LDAP + Cached Credentials, AutoFS, Network Booting Live CDs, etc. I simply now don't have to worry about the fundamentals being broken. I've only experienced a few Ubuntu bugs, but the time I've had to spend on them was far less than the compile time for one instance of Open Office.

[Arch]

I kinda don't get why everyone seems to just agree that it doesn't make sense to let a stock Ubuntu run against a fully customized Gentoo. Does seem unfair of course, but could give us a good idea how big real life differences between those system really are.

How about benchmarking a stock Ubuntu (--> mainstream system), a fully optimized Gentoo (--> tweaker's system), a recent version of Fedora (--> bleeding edge binary system) and a Debian system configured for maximum stability? (Not just in terms of cpu performance, but also things like power consumption, RAM usage, disk usage and so on). I guess some of the results would't be as significant as some people like to believe...

[monraaf]

Whatever I dictate the defaults to be when setting up my environment in .bashrc or specify them when building them on the build service.

Sorry if I wasn't clear. I was referring to the benchmark article, not to your post. After being told that:

• source based distros, like gentoo are the fastest
• it's confirmed that in most cases it's the fastest OS in the world
• and that In a fair comparison (all important software versions being the same), Gentoo likely would win in benchmarks.

I'm a bit underwhelmed by these benchmarks. So I'm wondering perhaps Michael didn't use the right CFLAGS. Also I'm wondering if the wise men from the Gentoo community (all 4 of them ) have reached their verdict and are declaring this benchmark flawed.