https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71614
David Edelsohn <dje at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dje at gcc dot gnu.org --- Comment #6 from David Edelsohn <dje at gcc dot gnu.org> --- The implementation of the __gcc_qmul multiplication routine is designed for the results to be within 2ulp of the correct result.