[LLVMbugs] [Bug 18347] New: Sparse switches lowered to an else-if chain

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=18347

            Bug ID: 18347
           Summary: Sparse switches lowered to an else-if chain
           Product: libraries
           Version: 3.3
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
          Assignee: unassignedbugs at nondot.org
          Reporter: jn at sirrida.de
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

in SelectionDAGBuilder::handleBTSplitSwitchCase
(lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp) seems to be a quirk which
can effectively create an else-if chain from a very sparse switch statement
instead of a balanced binary tree.

The reason is that a marginal single switch label is weighted too much causing
it to be split off; a possible solution is to weight it with 0. To do this one
can replace the assignments to LDensity und RDensity by:
    volatile double LDensity =
       (double)(LSize-1).roundToDouble() /
                           (LEnd - First + 1ULL).roundToDouble();
    volatile double RDensity =
      (double)(RSize-1).roundToDouble() /
                           (Last - RBegin + 1ULL).roundToDouble();

I replaced ?Size by (?Size-1).

Here is a more or less stupid test program (translate with "clang -O3 -S
test.cpp" and then look at test.s):

int main(int argc, char **argv) {
  switch (argc) {
    case 100: return 1;
    case 200: return 2;
    case 300: return 3;
    case 400: return 4;
    case 500: return 5;
    case 600: return 6;
    case 700: return 7;
    case 800: return 8;
    case 900: return 9;
    default: return 0;
    }
  }

As far as I can see this quirk is also present in trunk.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20131230/16dc7516/attachment.html>