> I hope that the Float arrays will be single precision.  Beyond 8-byte
> alignment and extra storage issues, there is a compelling performance issue
> as well.  The PPC supports a pipelined multiply-accumulate instruction for
> single precision numbers.  When methods for image/signal processing
> applications are translated to C it would be nice for the C-compiler to be
> able to make use of the mult-accum hardware.

But this is a lose for architectures with FPUs that handle doubles
as the normal situation.  I seem to recall that Intel chips use
80 bit floats and round/truncate to doubles and singles.  So to
specify singles means that some users would buy extra overhead to
see the full fruits of their processor's labor.

To be blunt, why should one CPU receive preferential attention?
Maybe this should be a compile time issue for the VM.

-- 
Christopher Oliver                     Traverse Internet
Systems Coordinator                    223 Grandview Pkwy, Suite 108
oliver at traverse.net                    Traverse City, Michigan, 49684
      Dedicated to perfecting the cluttered desktop metaphor.