[llvm-bugs] [Bug 35531] New: Wrong result after Loop Vectorization.

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=35531

            Bug ID: 35531
           Summary: Wrong result after Loop Vectorization.
           Product: new-bugs
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: ilia.taraban at intel.com
                CC: llvm-bugs at lists.llvm.org

This test fails with wrong result after Loop Vectorization:

================= nice.c ==============
unsigned int i = 0;
int main ()
{
    unsigned int m = i;
    unsigned int k = 1;
    for (i = 0; i <= 63; ++i) 
    {
        m &= k;
        m++;
    }
    printf("%u\n", m);
    return 0;
}
=======================================

>>> clang -v
clang version 6.0.0 (trunk 319765)
Target: x86_64-unknown-linux-gnu
Thread model: posix
...


>>> clang -O0 -o nice.exe nice.c
>> ./nice.exe
2

>>> clang -O2 -o nice.exe nice.c
>>> ./nice.exe
0

>>> clang -O2 -o nice.exe nice.c -mllvm -opt-bisect-limit=62 && ./nice.exe
...
BISECT: running pass (61) Rotate Loops on loop
BISECT: running pass (62) Loop Distribution on function (main)
BISECT: NOT running pass (63) Loop Vectorization on function (main)
BISECT: NOT running pass (64) Loop Load Elimination on function (main)
...
2

>>> clang -O2 -o nice.exe nice.c -mllvm -opt-bisect-limit=63 && ./nice.exe
...
BISECT: running pass (62) Loop Distribution on function (main)
BISECT: running pass (63) Loop Vectorization on function (main)
BISECT: NOT running pass (64) Loop Load Elimination on function (main)
BISECT: NOT running pass (65) Combine redundant instructions on function (main)
...
0

Let's look at IR before after Loop Vectorization:
============ nice.after.ll ============
...
  %9 = extractelement <16 x i1> %bin.rdx6, i32 0
  %10 = zext i1 %9 to i32; used i1, but variable m at the end of for can be
only 1 or 2.
  %cmp.n = icmp eq i32 64, 64
  br i1 %cmp.n, label %for.end, label %scalar.ph
...
for.end:                                          ; preds = %middle.block,
%for.body
  %inc.lcssa = phi i32 [ %inc, %for.body ], [ %10, %middle.block ]; result is
in %10
  store i32 64, i32* @i, align 4, !tbaa !2
  %call = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x
i8], [4 x i8]* @.str, i64 0, i64 0), i32 %inc.lcssa)
  ret i32 0
}
=======================================

-- 
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/20171205/c7027ae8/attachment.html>