This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: Wish for 2002 ...
James Antill <james@and.org> writes:
> Major snip of the Cc lines...
>
> tb@becket.net (Thomas Bushnell, BSG) writes:
>
> > Linus claimed to have something approaching "hard data" for the
> > proposition that adding functions to the library is inherently bad,
> > but he seems to have begged off of the request to tell me where I
> > could read more about it.
>
> Well it's fairly easy to do a simple benchmark of how much time glibc
> is taking to startup, take two /bin/true programs one in C and calls
> glibc exit() (linked static and strip'd) and the other in nasm from
> http://www.muppetlabs.com/~breadbox/software/tiny/teensy.html
Right, but we don't know whether this is because it has more
functions.
So, add two small functions to your library, and see what happens to
the startup cost; that's the relevant metric.
And moreover, if the actual problem is that it's slower to mmap a
large file than a small one, that's a kernel bug; there is no inherent
reason why it should be any slower.