Gentoo Archives: gentoo-user

From: Paul Hartman <paul.hartman+gentoo@×××××.com>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] A tale of computing thud and blunder
Date: Mon, 19 Dec 2011 16:39:28
Message-Id: CAEH5T2O1Xd=46G+-K4mMYEVdDTpc2EdnuRX903_L_mLHFUO5fQ@mail.gmail.com
In Reply to: Re: [gentoo-user] A tale of computing thud and blunder by Dale
1 On Mon, Dec 19, 2011 at 12:15 AM, Dale <rdalek1967@×××××.com> wrote:
2 > Pandu Poluan wrote:
3 >>
4 >>
5 >> Kind of like what I always do when I switch from -march=nocona to
6 >> -march=native. (Usually I use -march=nocona to ensure seamless VM migration
7 >> on my XenServer-equipped boxen, but for some VMs, i.e., those requiring me
8 >> to wring out every last drop of performance, I go native.)
9 >>
10 >> That said, if you want to experience fully the "GCC Graphite"
11 >> optimizations, you'll also want to do emerge -ev ;-)
12 >>
13 >> Rgds,
14 >>
15 >
16 >
17 > Is Graphite worthwhile on a desktop system or is it better suited for
18 > servers or both?  I found this but still not sure what it is intended for:
19 >
20 > http://gcc.gnu.org/wiki/Graphite/4.5
21 >
22 > Are there any reasons to leave this be for a while?  You know, bugs or
23 > packages that don't work with it?
24
25 I used it for @world and ran into a couple times when things were
26 broken with it enabled... Don't really remember specifics. I didn't
27 notice any perceivable speed difference, so I stopped using it just to
28 simplify things in the future. :)
29
30 In general,I have changed my CFLAGS to be less aggressive over the
31 past years. I started with the most aggressive optimizations, and
32 using -march=whatever... At some point I researched compile times for
33 different optimizations and realized some packages were taking tens of
34 minutes longer to compile! Surely whatever potential speed increase of
35 -O3 vs -O2 was not going to give me tens of minutes cumulatively. If
36 there is a package that really benefits from some specific
37 optimization (and doesn't already set those flags itself in the
38 ebuild), I'll set the CFLAGS for that individual package but leave the
39 rest of world very simplistic.
40
41 Same with -march, I've gone to using -mtune instead, after having to
42 move hard drives into a system with different CPU feature set. And for
43 virtual machines (in hosts I don't control, especially), they could be
44 migrated to some other machine without my knowledge, and i don't want
45 it to stop working if that happens.