Gentoo Archives: gentoo-dev

From: Paul de Vrieze <pauldv@g.o>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] Do we want optimal performance?
Date: Wed, 08 Sep 2004 19:54:35
Message-Id: 200409082154.31736.pauldv@gentoo.org
In Reply to: Re: [gentoo-dev] Do we want optimal performance? by Klavs Klavsen
1 On Wednesday 08 September 2004 21:11, Klavs Klavsen wrote:
2 > Chris Gianelloni said:
3 > [SNIP]
4 >
5 > > Gentoo is not all about performance. While many of our users want to
6 > > squeeze the every drop of performance out of their systems, many use
7 > > Gentoo for any number of other reasons such as our philosophy, our
8 > > community, the manageability of portage, or even because they think
9 > > Larry the Cow just owns.
10 >
11 > hehe. I totally agree. I choose Gentoo for the flexibility, for instance
12 > in getting the programversions I want to use - except I'm sad that
13 > gcc-versions get phased out too quickly IMHO, so I can't easily choose to
14 > stay with an "old" gcc like 3.2.2 - without running into packages I
15 > suddenly can't upgrade automatically, because they depend on a newer GCC,
16 > which I can see no reason for them to do.
17
18 Most applications should not explicitly depend on a specific version of the
19 compiler. There are however some serious bugs in certain compiler versions
20 that show up in certain complex applications (like openoffice). Those bugs
21 are sometimes a reason to stop support for that particular version (although
22 not using -march=pentium4 stops most 3.2.x family problems)
23 >
24 > You are probably right - especially when I'm told that gcc-3.5 has great
25 > profiling capabilities (GIMPLE - whatever that is :) - which I agree would
26 > be the better solution (so people can easily optimize their machine -
27 > doing profiles for their usage).
28
29 According to wikipedia and the gcc website it is an intermediate machine
30 independent language that is used for optimization.
31
32 > The idea would ofcourse be that, only the "obvious" programs would be
33 > tested - but if profiling were implemented/possible with gcc-3.5 and
34 > portage easily - I'm fairly certain that would be of more value (would
35 > that also help select the right CFLAGS ?)
36
37 Easy profiling mainly shows where bottlenecks are (not the subtle ones that
38 involve the cpu cache for example, for that you might want to try the
39 valgrind cachegrind module)
40
41 Paul
42
43 --
44 Paul de Vrieze
45 Gentoo Developer
46 Mail: pauldv@g.o
47 Homepage: http://www.devrieze.net