On 10/4/10, Wartan Hachaturow <wartan.hachaturow@gmail.com> wrote:
The same reasoning as before applies -- why would you want to do
this?
What are the benefits compared to multiarch ppc32/ppc64?
If you are using double precision floating point, the speed of a
program compiled 32bit on a 64bit machine is much much slower (maybe
1/10). The 64bit kernel supports 64bit operation Only if the program
is compile 64 bit ! Moreover with any app there are library calls.
However if compile 64bit, with altivec support, won't run on 32bit
machine. But with 32bit and altivec will run faster on both but
floating point not faster on on 64bit machine. But other things faster
on both.
best to try building and testing with various options to see.