https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123225
--- Comment #7 from vekumar at gcc dot gnu.org --- Do you have cases where uncounted loop vectorization is helpful?? Looks like real world benchmarks like Xalan, the string length is very small. Also, there is not much happening. Make sense for targets that could elide alignment checks (using non-faulting loads).
