[llvm] [X86][LoopVectorize] Enable MaximizeBandwidth by default on X86 with generic split cost floor (PR #201666)

[Sumukh J Bharadwaj via llvm-commits]

================
@@ -140,6 +140,11 @@ class X86TTIImpl final : public BasicTTIImplBase<X86TTIImpl> {
   TypeSize
   getRegisterBitWidth(TargetTransformInfo::RegisterKind K) const override;
   unsigned getLoadStoreVecRegBitWidth(unsigned AS) const override;
+  bool shouldMaximizeVectorBandwidth(
+      TargetTransformInfo::RegisterKind K) const override {
+    assert(K != TargetTransformInfo::RGK_Scalar);
----------------
amd-subharad wrote:

Thanks, will fix

https://github.com/llvm/llvm-project/pull/201666