Hi all,
On Friday 21 December 2007 20:08:32 John Griessen wrote:
> René Rebe wrote:
> > On Wednesday 19 December 2007 18:19:19 John Griessen wrote:
> is shared compiling across non-homogenous machines easy?
> >
> > If you want to utiize distcc or icecream you have to use the same
> > T2 tree to build the identical cross compiler on the other "nodes" -
> > or NFS mount the t2 cross compiler.
>
>
> Hmm... does the NFS mount method still give good performance?
> Does it transfer big enough chunks of work to help?
The NFS (or other networked FS) would just be to use the same
cross compiler binary. It usually will remain cached. This cross
compiler will not load many single files via the network, but
receive those preprocessed source stream via distcc or icecream
etc. So in this setup the performance should be pretty good.
> What's a compile time on one 2.5GHz computer with 0.5GB DDR memory
> for ARM with a small set of programs to be a gateway?
With uclibc and some minimal packages the compile time for
such as ARM target is just somewhere between 0.5 to 2 hours.
Mostly depending whether you choose glibc or the lightweight
uClibC and how many additional packages you select.
Even with basic X for e.g. a PDA compile time rarely increases
over 3h or so.
Also note that t2 only build newly selected packages or single
packages you ask it to rebuild. So when you just changes a
patch or other e.g. config files / scripts it will usually just
recompress the rootfs in some minutes - AND NOT rebuild
everything .-)
Yours,
--
René Rebe - ExactCODE GmbH - Europe, Germany, Berlin
Geschäftsführer: Susanne Klaus, René Rebe
Sitz: Berlin, Amtsgericht Charlottenburg HRB 105 123 B
USt-IdNr.: DE251602478
http://exactcode.de | http://t2-project.org | http://rene.rebe.name
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
lists@xxxxxxxxxxxxxx with a subject of: unsubscribe t2
|
