https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84547
--- Comment #2 from Richard Earnshaw <rearnsha at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #1) > Yes int128 (or rather double wide register) shifts are not optimized that > well. They are not used by many people even. I wonder if there is a way to > use vector instructions to do them. Even if there were (and I'm not aware of any), moving values into and out of the vector unit can be expensive so this would likely be a very poor solution.
