https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85544
--- Comment #12 from Thomas Koenig <tkoenig at gcc dot gnu.org> --- (In reply to Harald Anlauf from comment #10) > Handling positive powers of 2 should be straightforward: > > The condition is sth. like > > if (v > 1 && (v & (v-1) == 0)) > > and the exponent is derived by bit-counting the number of ones in (v-1)... ... which is something I (currently) do not know to do in the front end. Let's leave that as some optimization potential (it should also be possible to optimize (-2)**n and so on) for a later date. Otherwise: Patches welcome (as always :-)