[OLPC-devel] Re: [olpc-software] Federico Mena-Quintero's recent memory work.

[Jim Gettys]

He identified another place where people are systematically sloppy and
for which there is low hanging fruit.  I like low hanging fruit: it
tastes good ;-).  It will be interesting to see how much it saves, and
we may want to systematically examine all the libraries we depend on and
clean them up.

It does also beg the question of whether the current Linux shared
library scheme is a good one; some of the other ones don't have that
overhead, as the mail thread points out.
                      Regards,
                        - Jim


On Sun, 2006-04-23 at 18:17 -0400, Ivan Krstic wrote:
> Jim Gettys wrote:
> > http://primates.ximian.com/~federico/news.html
> 
> I don't know if you mean his work in general (which is excellent), or if
> you're specifically referring to the last entry about non-shared dirty
> lib data. That entry is easy to misread: he didn't cut 16MB out of
> memory requirements, nor did he find a way to. A lot of that 16MB is
> relocation data and can't be reclaimed by simply const'ing variables,
> and even variables that can be const'ed won't necessarily translate to
> direct memory use reductions.
> 
> In other words, this post is a best practices reminder, but don't expect
> memory requirements to drop by 16MB anytime soon.
> 
-- 
Jim Gettys
One Laptop Per Child