[llvm-bugs] [Bug 27643] New: [LV] Loop vectorizer does not respect vectorize_width pragma at -Os/-Oz

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=27643

            Bug ID: 27643
           Summary: [LV] Loop vectorizer does not respect vectorize_width
                    pragma at -Os/-Oz
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Loop Optimizer
          Assignee: unassignedbugs at nondot.org
          Reporter: mssimpso at codeaurora.org
                CC: llvm-bugs at lists.llvm.org, mcrosier at codeaurora.org
    Classification: Unclassified

When optimizing for size, the loop vectorizer chooses to not vectorize the
simple loop below due to the required run-time checks.

void f(int *a, int *b, int n) {
  for (int i = 0; i < n; ++i)
    a[i] = b[i];
}

Comments in the source suggest that vectorization can be forced with pragmas.
"#pragma clang loop vectorize(enable)" does indeed force vectorization, but
"#pragma clang loop vectorize_width(4)" does not. Specifying the width is
supposed to implicitly enable vectorization.

-- 
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/20160504/f404213f/attachment.html>