https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126964
--- Comment #2 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> --- Created attachment 65378 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=65378&action=edit A patch to test with actual benchmark I downloaded the modern version of tonto and could not reproduce. So I proceeded with a hunch that unnecessary copy in/out cycles were burning. Did some profiling. The attached patch is a lot of speculation on my part. If someone would please test and see if this fixes it?
