[comp.arch] RISC registers

henry@utzoo.uucp (Henry Spencer) (05/08/88)

> ... I still think that register windows are a poor design 
> choice in most cases...

Depends on what you are doing.  They appear to win heavily when function
calls are frequent and relatively unpredictable at compile time, as in
languages like C++ and Smalltalk.  I agree that their utility is much more
debatable for more predictable languages like C and Fortran.

> ...I wish that the chip real estate devoted to register windows would
> be devoted to a smaller number (32 say) all purpose 64 bit registers
> with better floating point support.

Surely they should be 80 bits or thereabouts, to support IEEE floating
point *properly*?
-- 
NASA is to spaceflight as            |  Henry Spencer @ U of Toronto Zoology
the Post Office is to mail.          | {ihnp4,decvax,uunet!mnetor}!utzoo!henry