Ive been hearing that slackware is the fastest linux distro out there for a while now. Now people are saying that Gentoo is the fastest. So I did some research..
Apparently (sorry if everyone else knows this, im always last to things...) gentoo optimizes all of its packages during the install (i think) so its kinda unfair to say that gentoo is faster than slackware right?
So lets say that you have 2 identical systems
1 with gentoo, 1 with slackware, except slackware has had all of its packages optimized. Oh and you do all the other optimizations (kernel recompile, dma etc). Which system would be faster?

Oh and is there a way to optimize all the packages without doing each one by hand?

As far as i can understand all Linux distros can be tailored to have pretty much the same speed. This is done by optimixing the kernel and stopping all un-needed services. What do you mena by package optimizations? In slackware most software you install you will have to compile first, which means it will be optimized for your system.
Anyone plz correct me if i'm wrong here..
-NSKL

Umm ... the basic stuff that comes with slack is
pre-compiled, it just gets untarred and configured,
whereas gentoo might be a source distro (? never seen it)

Anyway ... gcc by default generates 386 instructions,
but I reckon that most comps running Linux today can do
better than that ;)

Of course you are right about the services and the
startup-sequence :}

Cheers,
Tink

From the logic behind Gentoo, everything optimized to the hilt for your architecture, yeah it sounds like it'll end up being faster than Slack.

So what.

100Mph off of a cliff is still off of a cliff. Gentoo loves to do releases against RC kernels, yesterday's glibc, an hour ago's compiler. Patrick is conservative in how he packages things, and does it for i386, the slowest optimization. The big argument for Slack has always been that it is stable, the most Unix-like distro, and runs well on anything from a 386 to a P4.

It just happened to get the reputation for being rather quick because every other distro on the planet bloated as the hardware could handle it.

Also, all of that manic optimization for architecture really doesn't matter as much as say... 5 fewer processes running in "top".

Still, the fastest distro I've found is LFS... 11 processes running at startup, compared to Slackware's 33 and Redhat's 100.

Cheers,

Finegan

Since we are talking about services..

Is it possible to start a few required services for sound and graphical support, program X to startup, and once started have the workstation start all the remaining services.

It may seem silly, but honestly it would allow a user to get to the desktop much faster and start working if s/he did not have to wait so long for various services to start up.

I'm not a big X-junkie these days... actually I'm becoming one. rc.4 is usually reserved as a sort of roll-your-own init structure, its never the default runlevel. I guess one of the things Linux isn't really built to do is start up quickly as much as just start up without going kazoo... I guess you could put the call to start X sooner: for instance here's a snippet from my /etc/rc.d/rc5.d:

[root@nimble rc5.d]# ls
K03rhnsd K35vncserver K74ypxfrd S12syslog S28autofs S85gpm
K20nfs K45arpwatch K96irda S13portmap S45pcmcia S90crond
K20rstatd K46radvd S05kudzu S14nfslock S55sshd S90xfs
K20rusersd K50snmpd S08ipchains S17keytable S56rawdevices S95anacron
K20rwalld K65identd S08iptables S20random S56xinetd S95atd
K20rwhod K74ntpd S09isdn S25netfs S60lpd S99local
K34yppasswdd K74ypserv S10network S26apmd S80sendmail

The lower the S number, the sooner that they start, the K's are for kills and everything in there is just a symlink to a script in /etc/rc.d/init.d, so you could have xinitd called sooner than the rest of them. I've fiddled about with them a bit here and there to roll what I want and occasionally came out with conflicts, but its easy to debug.

Cheers,

Finegan

Finegan Thanks for the contribution, but since this is the Slackware forum, and I'm going to assume the question was in regards to a Slackware install, and Slackware uses bsd style init instead SysV, the above information is, um, completely and totally wrong.

For starters, runlevel 4 is the default X runlevel in slack (Runlevel 5 is the unused one.).

The best place to experiment with changing the order things are run in Slack with regards to demons starting before X would be in /etc/inittab. notice that rc.M is run before rc.4... You might want to place the line for rc.4 right before rc.M. Be sure to create a boot disk first, however, and let us know how it goes Alterntively, you might want to intergrate the contents of rc.4 file (not a directory) into rc.M so they execute *after* all the network stuff is setup, but before the slow demons are started (like Sendmail and LPD)

Of course, having posted that, I notice that Stephanie is an MDK user,,, <sigh> some days, it's just nor worth crawling out of bed.

Interesting, I never realized Slack used 4 instead of 5 as the default network and xdm runlevel (all my slack boxes start at 3, half don't even have X). Sorry, Steph's question was sort of an in-continuation from an assorted pile of other posts.

This is what threw me off the other day, from a Slack 8.0 /etc/inittab:

# Runlevel 4 used to be for an X window only system, until we discovered
# that it throws init into a loop that keeps your load avg at least 1 all
# the time. Thus, there is now one getty opened on tty6. Hopefully no one
# will notice. ;^)
# It might not be bad to have one text console anyway, in case something
# happens to X.
x1:4:wait:/etc/rc.d/rc.4

So Slack's X runlevel has only 1 tty? that's wiggy.

Cheers,

Finegan