public inbox for gentoo-user@lists.gentoo.org
 help / color / mirror / Atom feed
From: Alan McKinnon <alan.mckinnon@gmail.com>
To: gentoo-user@lists.gentoo.org
Subject: Re: [gentoo-user] upgrading from kernel 2.6.24-rc6 to latest kernel
Date: Mon, 18 May 2009 20:55:28 +0200	[thread overview]
Message-ID: <200905182055.28710.alan.mckinnon@gmail.com> (raw)
In-Reply-To: <20090518175914.GA28829@linux1>

On Monday 18 May 2009 19:59:14 William Hubbs wrote:
> On Mon, May 18, 2009 at 07:39:48PM +0200, Alan McKinnon wrote:
> > On Monday 18 May 2009 19:12:22 William Hubbs wrote:
> > > Another difference is that, since you are compiling everything from
> > > source, with the correct CFLAGS and CXXFLAGS settings in make.conf, you
> > > can optimize the binaries you produce to take full advantage of your
> > > processor, which you can't do on a binary distro since everything is
> > > already compiled for you.
> >
> > Another word:
> >
> > That's ricing.
> >
> > Ok, that's two words. You get the idea.
>
>  Actually, I don't, because I don't know what you mean.  I was just
>  pointing out something about gentoo which  I think is different from a
>  binary distribution.  I don't know how much of a difference it makes,
>  but it is a feature they don't have.

The cases where one really does need to optimize the compiler for your cpu are 
very rare, and only really apply to old hardware. Back in the day they 
performed fine, but code available today tends to do more (keeping pace with 
cpu capability) so you need to tweak things to extract the best performance.

If using current packages on current hardware, the benefit is questionable and 
users are highly unlikely to notice much difference between good old i686 and 
-O9. With some apps it does make a difference - extensive floating point ops 
comes to mind - and one should take advantage of those cases. However, looking 
into the ebuilds of such packages usually reveals that, if the maintainer is 
any good, those optimizations are already present in the ebuild.

Years ago Gentoo had a problem with fanatical moronic users claiming their 
machines performed a brazillion time faster with -O9 and other such nonsense. 
This is called "ricing" - a word play on young Japanese males modifying 
Subarus for "performance" that actually reduces performance. Ever seen a 
Subaru with a gigantic whale-tail rear spoiler? That's ricing. These Gentoo 
users seem to have gone away to wherever the current fashion fad is. For a 
while they infested Ubuntu. $DEITY only knows where they are now, maybe 
they're using MacOS and annoying Steve

Actual benchmarks should that ricing does nothing beneficial for the average 
ricer.

Genuine analysis of Gentoo machines admined by someone who knows how to do it 
should that the machine can easily have only the features and software on it 
that the admin say it should have. Like LDAP - not everyone needs it. On a 
binary distro, if the maintainer supports it you usually get it too whether 
you like it or not. With Gentoo, USE is your friend. And this is gentoo's 
greatest strength - the ability to build something much closer to what you 
really want than is possible with a binary distro.

-- 
alan dot mckinnon at gmail dot com



  parent reply	other threads:[~2009-05-18 18:56 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-05-12  8:19 [gentoo-user] upgrading from kernel 2.6.24-rc6 to latest kernel bn
2009-05-13  3:30 ` Adam Carter
2009-05-13 20:14   ` bn
2009-05-13 20:10     ` Paul Hartman
2009-05-13 20:11     ` [gentoo-user] " Nikos Chantziaras
2009-05-13 20:27       ` bn
2009-05-17 15:46 ` [gentoo-user] " bn
2009-05-17 15:57   ` Sean
2009-05-17 16:18   ` Philip Webb
2009-05-17 17:32     ` Neil Bothwick
2009-05-17 18:59       ` Dale
2009-05-18 11:29       ` Philip Webb
2009-05-18 11:57         ` Neil Bothwick
2009-05-18 14:06           ` Philip Webb
2009-05-18 16:09             ` bn
2009-05-18 16:28               ` Philip Webb
2009-05-18 16:42                 ` bn
2009-05-18 17:12                   ` William Hubbs
2009-05-18 17:39                     ` Alan McKinnon
     [not found]                       ` <20090518175914.GA28829@linux1>
2009-05-18 18:55                         ` Alan McKinnon [this message]
2009-05-18 19:21                           ` Dale
2009-05-18 18:38                     ` Dale
2009-05-18 17:38                   ` Alan McKinnon
2009-05-18 18:12                   ` Philip Webb
2009-05-18 20:14                     ` bn
2009-05-18 22:15                       ` Alan McKinnon
2009-05-19 16:23                         ` bn

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200905182055.28710.alan.mckinnon@gmail.com \
    --to=alan.mckinnon@gmail.com \
    --cc=gentoo-user@lists.gentoo.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox