[plug] Gentoo vs [insert ur fav distro here]

[James Devenish]

In message <1056068119.4741.55.camel at rattus.Localdomain>
on Fri, Jun 20, 2003 at 08:15:20AM +0800, William Kenworthy wrote:
> gcc3.x was generally considered noticeably faster

!?!?!?!?!?

In message <1056068119.4741.55.camel at rattus.Localdomain>
on Fri, Jun 20, 2003 at 08:15:20AM +0800, William Kenworthy wrote:
> where it counts

I have only heard specifically about what it does during compilation of
C source. I use it for C source and C++ source and I have found that the
difference can be as much as a 1.5min compile becoming 7mins. With
larger projects, the difference is usually less (e.g. factor of two??)
but the impact is greater since the original time was of the order of
hours or days.

In message <1056068119.4741.55.camel at rattus.Localdomain>
on Fri, Jun 20, 2003 at 08:15:20AM +0800, William Kenworthy wrote:
> I have seen nothing about being sower.

You've caught me in the middle of not compiling anything, so I have no
examples of my own at the moment (I just tried one on a machine with gcc
2.95.2 and gcc 3.2.2, but it needs to be re`configure`ed for each gcc
version and I don't have the time). Trying somewhere else...

% gcc -v |& grep version
gcc version 2.95.3 20010125 (prerelease)
% time make build
...
make clean build  4.47s user 2.32s system 92% cpu 7.356 total
% make clean
% ...change path for gcc 3.2...
% gcc -v |& grep version
gcc version 3.2
% time make build
...
make clean build  6.21s user 2.85s system 92% cpu 9.833 total

(I repeated this three times, data not shown. This is a very tame, and
probably lame, example but is the best I can do at this time of the
morning.)

In message <1056068119.4741.55.camel at rattus.Localdomain>
on Fri, Jun 20, 2003 at 08:15:20AM +0800, William Kenworthy wrote:
> Have not heard this before: do you have a reference for it?

Sure, will get back to you.