[LLVMbugs] [Bug 13050] New: nested loops leave dead loop for degenerate case

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=13050

             Bug #: 13050
           Summary: nested loops leave dead loop for degenerate case
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: hfinkel at anl.gov
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified


/test-suite/SingleSource/Regression/C/sumarray2d.c contains the function:
int SumArray(int Array[][100], unsigned int NumI, unsigned int NumJ) {
  unsigned i, j;
  int Result = 0;

  for (i = 0; i < NumI; i++)
    for (j = 0; j < NumJ; j++)
      Result += Array[i][j];

  return Result;
}

compiling this with clang -O3 produces:

define i32 @SumArray([100 x i32]* nocapture %Array, i32 %NumI, i32 %NumJ)
nounwind uwtable readonly {
entry:
  %cmp12 = icmp eq i32 %NumI, 0
  br i1 %cmp12, label %for.end8, label %for.cond1.preheader.lr.ph

for.cond1.preheader.lr.ph:                        ; preds = %entry
  %cmp29 = icmp eq i32 %NumJ, 0
  br i1 %cmp29, label %for.inc6, label %for.body3.lr.ph.us

... (regular loop processing) ...

for.inc6:                                         ; preds = %for.inc6,
%for.cond1.preheader.lr.ph
  %i.013 = phi i32 [ %inc7, %for.inc6 ], [ 0, %for.cond1.preheader.lr.ph ]
  %inc7 = add i32 %i.013, 1
  %exitcond20 = icmp eq i32 %inc7, %NumI
  br i1 %exitcond20, label %for.end8, label %for.inc6

... (end block)

I don't see any reason for the %for.inc6 loop to exists, it does not seem to
have any effect. Nothing in CodeGen eliminates it, and so it ends up in the
executable.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.