Examine the inner loop

This is the innermost loop after strength reduction and induction variable elimination:

L113:
   ldd [%o0],%f2
   fmuld %f6,%f2,%f2
   ldd [%o1],%f4     ! uses res of prev instn
   faddd %f2,%f4,%f2 ! uses res of prev instn
   std %f2,[%o1]     ! uses res of prev instn
   add %o0,8,%o0
   cmp %o0,%o2
   blu L113
   add %o1,8,%o1     ! (delay slot)

• Register o0 points to B[k][j].
• Register o1 points to C[i][j].
• Register o2 points to C[i][500].
• Register f6 holds r.
• There is some freedom in the order of instructions...