# r0=temp
# r1=source ptr (unaligned)
# r2=dest ptr (aligned)
# r3=number of words to copy
lp:   lwr   r0, 3(r1)     # <-- this ends up having a false dependency on r0 from the previous iter
      lwl   r0, 0(r1)
      sw    r0, 0(r2)
      addiu r1, r1, 4
      addiu r2, r2, 4
      addiu r3, r3, -1
      bgtz  r3, lp