SPARC compiler optimisation

I need to generate frequent checksums on a SPARC machine, and I have
been looking at the code the sun optimizer produces for the code:


main()
{
        int block[1000];
        register int i;
        register int xs=0;


        for(i=999;i>0;i--)
xs|=*(block+i);


        return xs;
}




The best that -O4 can do is a 5 tick loop:


L77010:
ld [%i5],%l3
dec 4,%i5
cmp %i5,%i3
bgu L77010
or %i4,%l3,%i4


where %i5 is set to block+999 and %i3 is block




It is possible to write a 4 tick loop (20% saving) :


LMYLOOP:
ld [%i3+%i5],%l3
deccc 4,%i5
bcc LMYLOOP
or %i4,%l3,%i4


where %i5 is i , and %i3 is block.




My question is why can the optimiser not generate this 4 tick loop?
The setup is simpler, the registers are the variables and
not combinations of the variables...


Looking at other code, it looks like the compiler/optimiser never
uses the [%R+%R] addressing mode. Is there a reason for this?


Does any other compiler do this? (I dont want to install gcc just to
answer this question).


Please email as well as post answers.


Greg Wilkins ( Software Manager ) @ Highland Logic
              Snail : 46 Albany St, St Leonards, NSW, 2065, Australia
              Email : gregw@highland.oz.au
              Phone : (+61 2) 9066624 OR (+61 48) 683490
              Fax : (+61 2) 4361540 OR (+61 48) 683474
--

Return to the comp.compilers page.
Search the comp.compilers archives again.