Index: llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp
===================================================================
--- llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp
+++ llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp
@@ -211,6 +211,11 @@
                      cl::desc("Verify this pass produces no dead code"),
                      cl::Hidden);
 
+static cl::opt<bool> ForceLowerGEP(
+    "force-lower-gep", cl::init(false),
+    cl::desc("Force to lower GEP during SeparateConstOffsetGEP pass"),
+    cl::Hidden);
+
 namespace {
 
 /// A helper class for separating a constant offset from a GEP index.
@@ -1167,6 +1172,9 @@
   if (DisableSeparateConstOffsetFromGEP)
     return false;
 
+  if (ForceLowerGEP)
+    LowerGEP = true;
+
   DL = &F.getParent()->getDataLayout();
   bool Changed = false;
   for (BasicBlock &B : F) {
Index: llvm/test/Transforms/SeparateConstOffsetFromGEP/split-gep-sub.ll
===================================================================
--- /dev/null
+++ llvm/test/Transforms/SeparateConstOffsetFromGEP/split-gep-sub.ll
@@ -0,0 +1,61 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -S -passes=separate-const-offset-from-gep -force-lower-gep < %s | FileCheck %s
+
+define void @test(ptr %p) {
+; CHECK-LABEL: @test(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = tail call i32 @foo()
+; CHECK-NEXT:    [[REM:%.*]] = srem i32 [[TMP0]], 5
+; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
+; CHECK:       for.body:
+; CHECK-NEXT:    [[K:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[COND_END:%.*]] ]
+; CHECK-NEXT:    [[MUL:%.*]] = mul nuw nsw i32 [[K]], 5
+; CHECK-NEXT:    [[SUB1:%.*]] = sub nsw i32 [[MUL]], [[REM]]
+; CHECK-NEXT:    [[CMP26:%.*]] = icmp ult i32 [[SUB1]], 512
+; CHECK-NEXT:    br i1 [[CMP26]], label [[COND_TRUE:%.*]], label [[COND_END]]
+; CHECK:       cond.true:
+; CHECK-NEXT:    [[SUB22:%.*]] = sext i32 [[SUB1]] to i64
+; CHECK-NEXT:    [[TMP1:%.*]] = ptrtoint ptr [[P:%.*]] to i64
+; CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[SUB22]], 2
+; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[TMP1]], [[TMP2]]
+; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[TMP3]], 2044
+; CHECK-NEXT:    [[TMP5:%.*]] = inttoptr i64 [[TMP4]] to ptr
+; CHECK-NEXT:    store float 1.000000e+00, ptr [[TMP5]], align 4
+; CHECK-NEXT:    br label [[COND_END]]
+; CHECK:       cond.end:
+; CHECK-NEXT:    [[INC]] = add nuw nsw i32 [[K]], 1
+; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i32 [[INC]], 100
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END:%.*]]
+; CHECK:       for.end:
+; CHECK-NEXT:    ret void
+;
+entry:
+  %0 = tail call i32 @foo()
+  %rem = srem i32 %0, 5
+  %add = add nsw i32 %rem , 511
+  br label %for.body
+
+for.body:
+  %k = phi i32 [ 0, %entry ], [ %inc, %cond.end ]
+  %mul = mul nuw nsw i32 %k, 5
+  %sub1 = sub nsw i32 %mul, %rem
+  %cmp26 = icmp ult i32 %sub1, 512
+  br i1 %cmp26, label %cond.true, label %cond.end
+
+cond.true:
+  %sub2 = sub nsw i32 %add, %mul
+  %idxprom = sext i32 %sub2 to i64
+  %arryidx = getelementptr inbounds float, ptr %p, i64 %idxprom
+  store float 1.0, ptr %arryidx, align 4
+  br label %cond.end
+
+cond.end:
+  %inc = add nuw nsw i32 %k, 1
+  %exitcond = icmp ne i32 %inc, 100
+  br i1 %exitcond, label %for.body, label %for.end
+
+for.end:
+  ret void
+}
+
+declare i32 @foo()