diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -3901,7 +3901,8 @@
     if (auto *SI = dyn_cast_or_null<StoreInst>(I)) {
       // Store instruction with index and scale costs 2 Uops.
       // Check the preceding GEP to identify non-const indices.
-      if (auto *GEP = dyn_cast<GetElementPtrInst>(SI->getPointerOperand())) {
+      if (auto *GEP = dyn_cast<GetElementPtrInst>(
+              SI->getPointerOperand()->stripPointerCasts())) {
         if (!all_of(GEP->indices(), [](Value *V) { return isa<Constant>(V); }))
           return TTI::TCC_Basic * 2;
       }
diff --git a/llvm/test/Transforms/LoopUnroll/X86/store_cost.ll b/llvm/test/Transforms/LoopUnroll/X86/store_cost.ll
--- a/llvm/test/Transforms/LoopUnroll/X86/store_cost.ll
+++ b/llvm/test/Transforms/LoopUnroll/X86/store_cost.ll
@@ -46,7 +46,9 @@
   store i32 %val4, i32* %xptr4
   %val5 = add i32 %counter, 10
   %xptr5 = getelementptr [1024 x i32], [1024 x i32]* %x05, i32 0, i32 %counter
-  store i32 %val5, i32* %xptr5
+  %bc = bitcast i32* %xptr5 to i16*
+  %val5trunc = trunc i32 %val5 to i16
+  store i16 %val5trunc, i16* %bc
   br label %loop.inc
 
 loop.inc: