[llvm] r234845 - [SCEV] Strengthen SCEVExpander::isHighCostExpansion.

[Sanjoy Das]

Author: sanjoy
Date: Mon Apr 13 22:20:32 2015
New Revision: 234845

URL: http://llvm.org/viewvc/llvm-project?rev=234845&view=rev
Log:
[SCEV] Strengthen SCEVExpander::isHighCostExpansion.

Summary:

Teach `isHighCostExpansion` to consider divisions by power-of-two
constants as cheap and add a test case.  This change is needed for a new
user of `isHighCostExpansion` that will be added in a subsequent change.

Depends on D8995.

Reviewers: atrick

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D8993

Added:
    llvm/trunk/test/Transforms/IndVarSimplify/lftr-udiv-tripcount.ll
Modified:
    llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp

Modified: llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp?rev=234845&r1=234844&r2=234845&view=diff
==============================================================================
--- llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp (original)
+++ llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp Mon Apr 13 22:20:32 2015
@@ -1810,12 +1810,23 @@ bool SCEVExpander::isHighCostExpansionHe
   if (!Processed.insert(S).second)
     return false;
 
-  // If the backedge-taken count is a UDiv, it's very likely a UDiv that
-  // ScalarEvolution's HowFarToZero or HowManyLessThans produced to compute a
-  // precise expression, rather than a UDiv from the user's code. If we can't
-  // find a UDiv in the code with some simple searching, assume the former and
-  // forego rewriting the loop.
-  if (isa<SCEVUDivExpr>(S)) {
+  if (auto *UDivExpr = dyn_cast<SCEVUDivExpr>(S)) {
+    // If the divisor is a power of two and the SCEV type fits in a native
+    // integer, consider the divison cheap irrespective of whether it occurs in
+    // the user code since it can be lowered into a right shift.
+    if (auto *SC = dyn_cast<SCEVConstant>(UDivExpr->getRHS()))
+      if (SC->getValue()->getValue().isPowerOf2()) {
+        const DataLayout &DL =
+            L->getHeader()->getParent()->getParent()->getDataLayout();
+        unsigned Width = cast<IntegerType>(UDivExpr->getType())->getBitWidth();
+        return DL.isIllegalInteger(Width);
+      }
+
+    // UDivExpr is very likely a UDiv that ScalarEvolution's HowFarToZero or
+    // HowManyLessThans produced to compute a precise expression, rather than a
+    // UDiv from the user's code. If we can't find a UDiv in the code with some
+    // simple searching, assume the former consider UDivExpr expensive to
+    // compute.
     BasicBlock *ExitingBB = L->getExitingBlock();
     if (!ExitingBB)
       return true;

Added: llvm/trunk/test/Transforms/IndVarSimplify/lftr-udiv-tripcount.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/IndVarSimplify/lftr-udiv-tripcount.ll?rev=234845&view=auto
==============================================================================
--- llvm/trunk/test/Transforms/IndVarSimplify/lftr-udiv-tripcount.ll (added)
+++ llvm/trunk/test/Transforms/IndVarSimplify/lftr-udiv-tripcount.ll Mon Apr 13 22:20:32 2015
@@ -0,0 +1,29 @@
+; RUN: opt < %s -indvars -S | FileCheck %s
+
+target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
+
+; It is okay to do LFTR on this loop even though the trip count is a
+; division because in this case the division can be optimized to a
+; shift.
+
+define void @foo(i8* %a, i8 %n) nounwind uwtable ssp {
+; CHECK-LABEL: @foo(
+ entry:
+  %e = icmp sgt i8 %n, 3
+  br i1 %e, label %loop, label %exit
+
+ loop:
+; CHECK-LABEL: loop:
+  %i = phi i8 [ 0, %entry ], [ %i.inc, %loop ]
+  %i1 = phi i8 [ 0, %entry ], [ %i1.inc, %loop ]
+  %i.inc = add nsw i8 %i, 4
+  %i1.inc = add i8 %i1, 1
+  store volatile i8 0, i8* %a
+  %c = icmp slt i8 %i, %n
+; CHECK-LABEL:  %exitcond = icmp ne i8 %i1.inc
+  br i1 %c, label %loop, label %exit
+
+ exit:
+; CHECK-LABEL: exit:
+  ret void
+}