[llvm] [SLP][AMDGPU] Vectorize operands of non-trivially-vectorizable intrinsic calls (PR #189784)

[Alexey Bataev via llvm-commits]

================
@@ -29477,6 +29510,22 @@ bool SLPVectorizerPass::vectorizeChainsInBlock(BasicBlock *BB, BoUpSLP &R) {
       PostProcessCmps.insert(cast<CmpInst>(&*It));
   }
 
+  DenseMap<Intrinsic::ID, SmallSetVector<Value *, 4>> IntrinsicSeedOps;
----------------
alexey-bataev wrote:

Use SmallMapVector, otherwise iteration over unordered container will cause non-determinism

https://github.com/llvm/llvm-project/pull/189784