[PATCH] D20315: [LV] For some induction variables, use vector phis instead of widening the scalar in the loop body

[Michael Kuperstein via llvm-commits]

mkuper added a comment.

Right, I'll add an explicit test (the test in induction_plus is *sort of* that, but not quite), thanks David!

Regarding the extra movdqa - I think the copy may be necessary.
The problem is that the loop body needs both to modify the current IV (because of two-address instructions) and keep it so that it can generate the new IV.

GCC does something similar:

  jmp	.L3
  [...]
  .L5:
  	movdqa	%xmm4, %xmm1
  .L3:
  	movdqa	%xmm1, %xmm4
  	pxor	%xmm6, %xmm1
  	movdqa	%xmm5, %xmm2
  	addl	$1, %eax
  	cmpl	$250, %eax
  	paddd	%xmm7, %xmm4
  	pcmpgtd	%xmm1, %xmm2
  	movdqa	%xmm1, %xmm3
  	punpckhdq	%xmm2, %xmm1
  	punpckldq	%xmm2, %xmm3
  	paddq	%xmm3, %xmm0
  	paddq	%xmm1, %xmm0
  	jne	.L5

The difference is that the loop is rotated, so there is no copy on the first iteration, but other than that, we still have the same two movdqas per iteration.

In any case, the extra mov disappears once we have AVX and three-address instructions, so we no longer update the current IV destructively.


http://reviews.llvm.org/D20315