[SCEV] Add a threshold to restrict number of mul operands to be inlined into SCEV

This is to avoid inlining too many multiplication operands into a SCEV, which could
take exponential time in the worst case.

Reviewers: Sanjoy Das, Mehdi Amini, Michael Zolotukhin

Differential Revision: https://reviews.llvm.org/D25794

llvm-svn: 284784

diff --git a/llvm/lib/Analysis/ScalarEvolution.cpp b/llvm/lib/Analysis/ScalarEvolution.cpp
index 9fa0de1..8c6ddff 100644 (file)
--- a/llvm/lib/Analysis/ScalarEvolution.cpp
+++ b/llvm/lib/Analysis/ScalarEvolution.cpp
@@ -121,6 +121,11 @@ static cl::opt<bool>
                   cl::desc("Verify no dangling value in ScalarEvolution's "
                            "ExprValueMap (slow)"));
 
+static cl::opt<unsigned> MulOpsInlineThreshold(
+    "scev-mulops-inline-threshold", cl::Hidden,
+    cl::desc("Threshold for inlining multiplication operands into a SCEV"),
+    cl::init(1000));
+
 //===----------------------------------------------------------------------===//
 //                           SCEV class definitions
 //===----------------------------------------------------------------------===//
@@ -2516,6 +2521,8 @@ const SCEV *ScalarEvolution::getMulExpr(SmallVectorImpl<const SCEV *> &Ops,
   if (Idx < Ops.size()) {
     bool DeletedMul = false;
     while (const SCEVMulExpr *Mul = dyn_cast<SCEVMulExpr>(Ops[Idx])) {
+      if (Ops.size() > MulOpsInlineThreshold)
+        break;
       // If we have an mul, expand the mul operands onto the end of the operands
       // list.
       Ops.erase(Ops.begin()+Idx);

diff --git a/llvm/test/Analysis/ScalarEvolution/max-mulops-inline.ll b/llvm/test/Analysis/ScalarEvolution/max-mulops-inline.ll
new file mode 100644 (file)
index 0000000..c0dc6e0
--- /dev/null
+++ b/llvm/test/Analysis/ScalarEvolution/max-mulops-inline.ll
@@ -0,0 +1,29 @@
+; RUN: opt -analyze -scalar-evolution -scev-mulops-inline-threshold=1 < %s | FileCheck --check-prefix=CHECK1 %s
+; RUN: opt -analyze -scalar-evolution -scev-mulops-inline-threshold=10 < %s | FileCheck --check-prefix=CHECK10 %s
+
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+@a = local_unnamed_addr global i32 0, align 4
+@b = local_unnamed_addr global i32 0, align 4
+
+define i32 @main() local_unnamed_addr {
+
+; CHECK1: %mul.1 = mul nsw i32 %mul, %mul
+; CHECK1: -->  ((%a.promoted * %a.promoted) * (%a.promoted * %a.promoted))
+
+; CHECK10: %mul.1 = mul nsw i32 %mul, %mul
+; CHECK10: -->  (%a.promoted * %a.promoted * %a.promoted * %a.promoted)
+
+entry:
+  %a.promoted = load i32, i32* @a, align 4
+  %mul = mul nsw i32 %a.promoted, %a.promoted
+  %mul.1 = mul nsw i32 %mul, %mul
+  %mul.2 = mul nsw i32 %mul.1, %mul.1
+  %mul.3 = mul nsw i32 %mul.2, %mul.2
+  %mul.4 = mul nsw i32 %mul.3, %mul.3
+  %mul.5 = mul nsw i32 %mul.4, %mul.4
+  store i32 %mul.5, i32* @a, align 4
+  store i32 31, i32* @b, align 4
+  ret i32 0
+}