[fpu] FPU operations

[Damjan Lampret <damjanlampret@yahoo.com>]

Hi all,

I have a question about FP operations performed by FPU (I am adding FP
instructions to OR1K - only single precision at the moment). Anyway
I'll list some of the operations that would be nice to have and I hope
someone can comment if they are simple enough to be supported by FPU.

So basic operations that should be supported are:

- SF add (SF stands for single precision FP)
- SF sub (should this be done with neg+add combination?)
- SF mul
- SF compare
- conversion operations (should integer<->SF be done by basic FPU or is
there any other idea? Also don't forget if you want DF then we need
SF<->DF)

Extra operations (desired but not really required IMO):
- SF div
- SF mac (or you can call it madd. Anyway it would be nice to have this
one. It is easy to implement it if FP Add and Mul hardware is separate)
- SF sqrt
- SF neg (MIPS, SPARC & PPC have it - is it required by the standard?)
- SF abs (MIPS, SPARC & PPC have it - is it required by the standard?)

Instructions like floating point moves between FP registers or
load/stores to/from FP registers are performed directly over FP
register file (no processing required) so I didin't mentioned them
above (I only listed operations that FP calculation hardware should be
able to do).

I can list FP instructions of MIPS, SPARC and other processor
architectures if there is interest. 

regards,
Damjan


__________________________________________________
Do You Yahoo!?
Kick off your party with Yahoo! Invites.
http://invites.yahoo.com/