compile speed on lfs 5.0

Ken Moffat ken at
Fri May 7 13:19:31 PDT 2004

On Fri, 7 May 2004, mrnobby wrote:

> Unfortunately not, I got cocky and deleted it once I got LFS up and
> running. Now, it *is* only a 'breadboard' system so I could wipe it and
> start over. The LFS 5.0 gcc version is 3.3.1. I'll have to find out the
> RH 7.0 version. Or someone may know off hand.

 RH 7.0 was probably gcc-2.95.3.  I installed RH 6.0 the other week on
one box, and that has egcs (2.91.66).  Might use that for testing, but
I've still go to upgrade it to a 2.4 kernel.  Sometime they went to what
they called gcc-2.96 which was "interesting" (and broken in the initial
versions), but I assume that was after 7.0 (I was using mdk in those

 Now, gcc-2.95.3 is a rocket compared to any of the gcc-3.3 family.
But once you've completed pass 1 of binutils and bash, everything is
using the new compiler anyway so although the elapsed times for these
first two become much longer, everything else should be 'similar'.  My
understanding was that *everything* was taking a lot longer.

 'Similar' means that when you build from an up-to-date host, things
like info pages will be built, whereas on an older host they may be
skipped if makeinfo is too old.  Maybe some other work-arounds will or
won't come into play, but those should be minor in the overall scheme of

 The other thing that would slow an old box down is forgetting to use
swap. but I'm sure you've checked that in `top'.

 das eine Mal als Tragödie, das andere Mal als Farce

More information about the lfs-support mailing list