Build time information

Robert Connolly robert at linuxfromscratch.org
Fri Jan 13 20:57:17 PST 2006


Thank you for your report and contribution. I noticed there is a significant 
variation in times for packages in uclibc and glibc which have identical 
conditions, such as the kernel headers and cross tools, but I don't think 
variations in times can be avoided. I'll try to add SBU's and package/patch 
file sizes soon.

robert

On January 12, 2006 03:36 pm, Tha D0od wrote:
> I got to wondering how long it would take to build HLFS from the book after
> noticing the SBU information was missing.  It takes quite a fair amount of
> time especially when you're watching compiler messages scroll down the
> screen.
>
> Anyway, I don't know if anyone's done this specifically for HLFS yet, but
> here are my time stats for the temporary utility build for both glibc and
> uClibc (chapter 5).  I also threw in some 'du -sh' info, just for overkill.
> I also took the time to normalise the results to the first build of
> binutils using the 'real' time value output of the time command.
>
> The results are kinda interesing.  No tests were run, and the output for
> the time command is for all commands AFTER patching and configuring.  That
> means, including make, make install, ln -s <blah> <blah>, cp <blah> <blah>,
> etc.  All done on a first gen Centrino (1M cache) laptop clocked at 1.6GHz
> with two ttys open (lynx up on one of them displaying the respecitve book
> being worked on) booted off the LFS 6.1-3 live CD.



More information about the hlfs-dev mailing list