[llvm] [AMDGPU] Hoist readlane/readfirst through unary/binary operands (PR #129037)

[Jay Foad via llvm-commits]

================
@@ -481,6 +482,65 @@ bool GCNTTIImpl::simplifyDemandedLaneMaskArg(InstCombiner &IC,
   return false;
 }
 
+Instruction *GCNTTIImpl::hoistReadLaneThroughOperand(InstCombiner &IC,
+                                                     IntrinsicInst &II) const {
+  Instruction *Op = dyn_cast<Instruction>(II.getOperand(0));
+
+  // Only do this if both instructions are in the same block
+  // (so the exec mask won't change) and the readlane is the only user of its
+  // operand.
+  if (!Op || !Op->hasOneUser() || Op->getParent() != II.getParent())
+    return nullptr;
+
+  const bool IsReadLane = (II.getIntrinsicID() == Intrinsic::amdgcn_readlane);
+
+  // If this is a readlane, check that the second operand is a constant, or is
+  // defined before Op so we know it's safe to move this intrinsic higher.
+  Value *LaneID = nullptr;
+  if (IsReadLane) {
+    LaneID = II.getOperand(1);
+    // Check LaneID is available at Op, otherwise we can't move the readlane
+    // higher.
+    if (!IC.getDominatorTree().dominates(LaneID, Op))
+      return nullptr;
+  }
+
+  const auto DoIt = [&](unsigned OpIdx) -> Instruction * {
+    SmallVector<Value *, 2> Ops{Op->getOperand(OpIdx)};
+    if (IsReadLane)
+      Ops.push_back(LaneID);
+
+    // Make sure convergence tokens are preserved.
+    // TODO: CreateIntrinsic should allow directly copying bundles
+    SmallVector<OperandBundleDef, 2> OpBundles;
+    II.getOperandBundlesAsDefs(OpBundles);
+
+    CallInst *NewII =
+        IC.Builder.CreateCall(II.getCalledFunction(), Ops, OpBundles);
+
+    Instruction &NewOp = *Op->clone();
+    NewOp.setOperand(OpIdx, NewII);
+    return &NewOp;
+  };
+
+  // TODO: Are any operations more expensive on the SALU than VALU, and thus
+  //       need to be excluded here?
+
+  if (isa<UnaryOperator>(Op))
----------------
jayfoad wrote:

If you also test for `CastInst`, will this subsume #128494?

https://github.com/llvm/llvm-project/pull/129037