If this works well I can surely setup one aranym instance and then use a distcc cross instance on my dual proc Xeon box :-P That should produce some stellar build times for the packages that can handle distcc. Thanks, Mark On Fri, 2006-10-20 at 09:42 +0200, Bill Allombert wrote: > Dear Debian 68k porters, > > I successfully build pari under aranym using distcc and gcc-cross > running on the host. > > The full buildlog is available here: > <http://people.debian.org/~ballombe/misc/pari_2.3.1-2_m68k.build.distcc> > > It took 7h30, but there is still room for improving the set up > (due to a bug in aranym some files were compiled locally, which slowdown > the build a lot) > > It should be compared to the buildlog on hobbes: > <http://buildd.Debian.org/fetch.cgi?pkg=pari;ver=2.3.1-2;arch=m68k;stamp=1161007964> > > Of course it is possible to run distcc+gcc-cross on real 68k hardware, > which will be faster. The advantage of aranym is that you only need to > take care of one physical machine. > > There were an extra warning though: > cc -c -g -O3 -Wall -fomit-frame-pointer -fno-strict-aliasing -I. -I../src/headers -D_REENTRANT -o base1.o ../src/basemath/base1.c > ../src/basemath/base1.c: In function 'get_roots_for_M': > ../src/basemath/base1.c:1051: warning: implicit declaration of function 'log2' > ../src/basemath/base1.c:1051: warning: incompatible implicit declaration of built-in function 'log2' > > Note: I tested pari because: > - I maintain it so I know it well > - It includes inline asm code for better m68k support. > - It is known to trigger gcc optimisation bug. > - It includes a test-suite so we can detect compilation problems. > - It is moderately large. > > Cheers, > -- > Bill. <ballombe@debian.org> > > Imagine a large blue swirl here. > >
Attachment:
signature.asc
Description: This is a digitally signed message part