ppc64 builds

[David Woodhouse]

On Mon, 2007-03-19 at 02:51 -0400, Tom Lane wrote:
> AFAIK, not building a ppc64 version should not be a "bug", indeed it
> likely should be the default.  32-bit code runs faster than 64-bit code
> on that arch, and so the only reason to build 64-bit is if you really
> need access to more than 4Gb of address space.  There are apps that need
> that, but not all that many.
> 
> So my position is that libraries should generally be built in both
> flavors (since they can't predict which flavor of executable might want
> them) but applications should be built as 32-bit unless there's a good
> reason why they need a 64-bit address space.

Certainly I agree that we should _ship_ no executables in 64-bit mode
other than the ones which actually need it, like perhaps gdb.

However, we currently need a pure 64-bit buildroot for building the
64-bit stuff, so in practice we'll need to build _some_ 64-bit packages
that don't get shipped. It might well be easier just to build them all.

> (If you ask me, the interesting question here is why the other arches
> don't behave the same way.  Any reasonably competent hardware design
> should have the property that doubling the bit-volume of traffic to
> main memory has a penalty...)

We do it the other way round on i386 because i386 _isn't_ something we
could consider a "reasonably competent" design these days. It has too
few registers, and the benefit of switching to x86_64 mode and actually
having a sensible number of registers is more than sufficient to offset
the bloat of 64-bit code.

-- 
dwmw2