On Mon, 8 Nov 1999, Bill Rea wrote:
> I'm a bit red-faced on this one. I just tried it again and it doesn't.
> This is still a mystery to me. It would seems to me that for
> this type of code that having full access to the 64-bit instruction
> set of the UltraSPARC CPUS and running it on a 64-bit operating system
> would give you the best performance. But that doesn't seem to be the
> case.
>
As long as you talk only about floating point instructions, even the
gnu assembler uses all the 64-bit code you're likely to see. The only
extension the SPARC V9 instruction set has that would be of interest
is a 64-bit load and store; but the gnu assembler seems to alias the
assembly mnemonic for "ldd" and "std" to the V9 version automatically
(I guess so that older FPU code can take advantage of the new faster
instructions). Either that, or (more likely) the UltraSPARC doesn't
support the V8 load and store double at all, but uses its opcode for the
newer V9 version.
jasonp
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
