diff --git a/llvm/lib/Target/X86/X86PreTileConfig.cpp b/llvm/lib/Target/X86/X86PreTileConfig.cpp
--- a/llvm/lib/Target/X86/X86PreTileConfig.cpp
+++ b/llvm/lib/Target/X86/X86PreTileConfig.cpp
@@ -31,6 +31,7 @@
 #include "llvm/CodeGen/MachineFunctionPass.h"
 #include "llvm/CodeGen/MachineInstr.h"
 #include "llvm/CodeGen/MachineLoopInfo.h"
+#include "llvm/CodeGen/MachineModuleInfo.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/CodeGen/TargetInstrInfo.h"
@@ -40,10 +41,16 @@
 using namespace llvm;
 
 #define DEBUG_TYPE "tile-pre-config"
-#define REPORT_CONFIG_FAIL                                                     \
-  report_fatal_error(                                                          \
-      MF.getName() +                                                           \
-      ": Failed to config tile register, please define the shape earlier");
+
+static void emitErrorMsg(MachineFunction &MF) {
+  SmallString<32> Str;
+  Twine ErrorMsg =
+      MF.getName() +
+      ": Failed to config tile register, please define the shape earlier";
+  LLVMContext &Context = MF.getMMI().getModule()->getContext();
+  Context.emitError(ErrorMsg);
+  exit(1);
+}
 
 namespace {
 
@@ -303,11 +310,11 @@
   for (auto &I : ShapeBBs) {
     // TODO: We can hoist shapes across BBs here.
     if (BBVisitedInfo[I.first].HasAMXRegLiveIn)
-      REPORT_CONFIG_FAIL
+      emitErrorMsg(MF);
     if (BBVisitedInfo[I.first].FirstAMX &&
         BBVisitedInfo[I.first].FirstAMX < I.second.back() &&
         !hoistShapesInBB(I.first, I.second))
-      REPORT_CONFIG_FAIL
+      emitErrorMsg(MF);
     WorkList.push_back(I.first);
   }
   while (!WorkList.empty()) {
diff --git a/llvm/test/CodeGen/X86/AMX/amx_error.ll b/llvm/test/CodeGen/X86/AMX/amx_error.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/X86/AMX/amx_error.ll
@@ -0,0 +1,13 @@
+@row = dso_local global i16 8, align 2
+@col = dso_local global i16 8, align 2
+
+define dso_local void @add() {
+entry:
+  %t0 = load i16, ptr @row, align 2
+  %t1 = call x86_amx @llvm.x86.tilezero.internal(i16 %t0, i16 64)
+  %t2 = load i16, ptr @col, align 2
+  %t3 = call x86_amx @llvm.x86.tilezero.internal(i16 16, i16 %t2)
+  ret void
+}
+
+declare x86_amx @llvm.x86.tilezero.internal(i16, i16)