[Bug tree-optimization/124035] New: GCC does not support early inlining of memcpy of non-power-of 2

[tnfchris at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124035

            Bug ID: 124035
           Summary: GCC does not support early inlining of memcpy of
                    non-power-of 2
           Product: gcc
           Version: 16.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tnfchris at gcc dot gnu.org
  Target Milestone: ---

The following

void f(char *a, char *b, unsigned s)
{
  __builtin_memcpy (a, b, 4);
}

void g(char *a, char *b, unsigned s)
{
  __builtin_memcpy (a, b, 3);
}

f is inlined early but not g, which means if inside a loop we won't vectorize
like clang does.

GCC only seems to do early inlining when the copy amount if a power of two