🚀 go-pugleaf

In article <1991Apr10.131341.26357@b11.ingr.com> dclark@b11.ingr.com (Dave Clark) writes:
> Although I haven't (yet) seen an assembler that does
>optimizations, I have seen utilities that can optimize assembly
>code (part of an HOL compiler back-end).

	Actually, (somewhat) optimizing assemblers are by no means uncommon.
For example, the 680x0 assembler I use converts my lsl.l #1,Dn instruction
into add.l Dn,Dn for me (there are ones that do far more wide-ranging
optimizations, used as back-ends for compilers).

>One of the compilers I'm using actually does instruction reordering
>within the linker!

	That's the right place to do it!  If you have deep pipelines, being
forced to make worst-case assumptions at routine entry can be quite limiting.
By doing reorganization in (or after) the linker, you know all the pre-
conditions a routine will have to deal with (unless someone takes it's
address, of course - then you go back to worst-case, unless you have a VERY
smart and lucky linker).

	The RPM-40 backend team (which I was part of) made use of this
trick, since we had long pipelines and load-delays (plus load/store/bra
interlocks to worry about).  Libraries were supposed to be in effectively
source format.

-- 
Randell Jesup, Keeper of AmigaDos, Commodore Engineering.
{uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.commodore.com  BIX: rjesup  
Disclaimer: Nothing I say is anything other than my personal opinion.
Thus spake the Master Ninjei: "To program a million-line operating system
is easy, to change a man's temperament is more difficult."
(From "The Zen of Programming")  ;-)