Index: lib/Target/AArch64/AArch64Subtarget.h
===================================================================
--- lib/Target/AArch64/AArch64Subtarget.h
+++ lib/Target/AArch64/AArch64Subtarget.h
@@ -85,6 +85,7 @@
   bool HasArithmeticCbzFusion = false;
   bool DisableLatencySchedHeuristic = false;
   bool UseRSqrt = false;
+  bool UseUnalignedQuadwordStorePenalty = true;
   uint8_t MaxInterleaveFactor = 2;
   uint8_t VectorInsertExtractBaseCost = 3;
   uint16_t CacheLineSize = 0;
@@ -196,6 +197,9 @@
   bool hasArithmeticBccFusion() const { return HasArithmeticBccFusion; }
   bool hasArithmeticCbzFusion() const { return HasArithmeticCbzFusion; }
   bool useRSqrt() const { return UseRSqrt; }
+  bool useUnalignedQuadwordStorePenalty() const {
+    return UseUnalignedQuadwordStorePenalty;
+  }
   unsigned getMaxInterleaveFactor() const { return MaxInterleaveFactor; }
   unsigned getVectorInsertExtractBaseCost() const {
     return VectorInsertExtractBaseCost;
Index: lib/Target/AArch64/AArch64Subtarget.cpp
===================================================================
--- lib/Target/AArch64/AArch64Subtarget.cpp
+++ lib/Target/AArch64/AArch64Subtarget.cpp
@@ -71,6 +71,7 @@
     break;
   case Falkor:
     MaxInterleaveFactor = 4;
+    UseUnalignedQuadwordStorePenalty = false;
     break;
   case Kryo:
     MaxInterleaveFactor = 4;
@@ -79,6 +80,7 @@
     PrefetchDistance = 740;
     MinPrefetchStride = 1024;
     MaxPrefetchIterationsAhead = 11;
+    UseUnalignedQuadwordStorePenalty = false;
     break;
   case Vulcan:
     MaxInterleaveFactor = 4;
Index: lib/Target/AArch64/AArch64TargetTransformInfo.cpp
===================================================================
--- lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -467,7 +467,8 @@
                                     unsigned Alignment, unsigned AddressSpace) {
   std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Src);
 
-  if (Opcode == Instruction::Store && Src->isVectorTy() && Alignment != 16 &&
+  if (ST->useUnalignedQuadwordStorePenalty() && Opcode == Instruction::Store &&
+      Src->isVectorTy() && Alignment != 16 &&
       Src->getVectorElementType()->isIntegerTy(64)) {
     // Unaligned stores are extremely inefficient. We don't split
     // unaligned v2i64 stores because the negative impact that has shown in
Index: test/Analysis/CostModel/AArch64/falkor.ll
===================================================================
--- /dev/null
+++ test/Analysis/CostModel/AArch64/falkor.ll
@@ -0,0 +1,15 @@
+; RUN: opt < %s -cost-model -analyze -mcpu=falkor | FileCheck %s
+
+target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
+target triple = "aarch64--linux-gnu"
+
+; CHECK-LABEL: memoryOpCost
+define void @memoryOpCost(<2 x i64> %a, <2 x i64>* %ptr) {
+
+  ; Disable the unaligned quadword store penalty for falkor.
+  ;
+  ; CHECK: cost of 1 {{.*}} store <2 x i64> %a, <2 x i64>* %ptr, align 8
+  store <2 x i64> %a, <2 x i64>* %ptr, align 8
+
+  ret void
+}
Index: test/Analysis/CostModel/AArch64/kryo.ll
===================================================================
--- test/Analysis/CostModel/AArch64/kryo.ll
+++ test/Analysis/CostModel/AArch64/kryo.ll
@@ -24,3 +24,14 @@
 
     ret void
 }
+
+; CHECK-LABEL: memoryOpCost
+define void @memoryOpCost(<2 x i64> %a, <2 x i64>* %ptr) {
+
+  ; Disable the unaligned quadword store penalty for kryo.
+  ;
+  ; CHECK: cost of 1 {{.*}} store <2 x i64> %a, <2 x i64>* %ptr, align 8
+  store <2 x i64> %a, <2 x i64>* %ptr, align 8
+
+  ret void
+}