[PATCH] D116270: [AMDGPU] Enable divergence-driven XNOR selection

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec added a comment.

LGTM provided comments are fixed.



================
Comment at: llvm/include/llvm/CodeGen/TargetLowering.h:3293
 
+  virtual bool isReassocProfitable(SelectionDAG &DAG, SDValue N0,
+                                   SDValue N1) const {
----------------
foad wrote:
> Needs a proper descriptive comment.
Still needs comment.


================
Comment at: llvm/lib/Target/AMDGPU/SIISelLowering.cpp:12494
+                                                 SDValue N1) const {
+  if (N0.hasOneUse()) {
+    // Take care of the oportunity to keep N0 uniform
----------------
Early return instead.


================
Comment at: llvm/lib/Target/AMDGPU/SIISelLowering.cpp:12496
+    // Take care of the oportunity to keep N0 uniform
+    if (!(!N0->isDivergent() && N1->isDivergent()))
+      return true;
----------------
Demorgan this.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D116270/new/

https://reviews.llvm.org/D116270