[PATCH] D154314: [LV] Remove the reminder loop if we know the mask is always true

[David Sherwood via Phabricator via llvm-commits]

david-arm added inline comments.


================
Comment at: llvm/test/Transforms/LoopVectorize/AArch64/eliminate-tail-predication.ll:34
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1024, [[N_VEC]]
-; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
----------------
I'm guessing that InstCombine does not determine this is guaranteed to always be true? However, I thought that someone did work in the DAGCombiner that will replace this with

  br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]]

when vscale is known to be a power of 2? Are you hoping to benefit from eliminating the scalar tail in IR because it helps us to make better decisions later in the pipeline? I can imagine it's beneficial for LTO where the scalar tail could prevent inlining.

If I remember correctly one of the problems with folding away the icmp in InstCombine is that it doesn't have access to the TTI interface so we cannot query the target.


================
Comment at: llvm/test/Transforms/LoopVectorize/X86/constant-fold.ll:30
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i32 2, 2
-; CHECK-NEXT:    br i1 [[CMP_N]], label [[BB3:%.*]], label [[SCALAR_PH]]
----------------
For all the fixed-lenth vector tests this icmp will get replaced with "i1 true" by InstCombine so the scalar tail should get automatically deleted.


================
Comment at: llvm/test/Transforms/LoopVectorize/dont-fold-tail-for-divisible-TC.ll:4
 
 target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
 
----------------
Hmm, I know this is not introduced by your patch, but I don't think we should have tests with target-specifics in the top level LoopVectorize directory.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D154314/new/

https://reviews.llvm.org/D154314