I agree, the best is to speed up matrices at all, instead of creating a dangerous hack, just to speed up 2x2 Matrices.
Further I think it's natural to inherit (most) matrix classes from Matrix, so this should not be changed. bg, Johannes On 11.01.2013 21:47, Nils Bruin wrote: > sage/matrix$ grep _mutability *.pxd > matrix0.pxd: cdef sage.structure.mutability.Mutability _mutability > matrix0.pxd: cdef check_mutability(self) > matrix0.pxd: cdef check_bounds_and_mutability(self, Py_ssize_t i, > Py_ssize_t j) > matrix0.pxd: cdef check_row_bounds_and_mutability(self, Py_ssize_t > r1, Py_ssize_t r2) > matrix0.pxd: cdef check_column_bounds_and_mutability(self, > Py_ssize_t c1, Py_ssize_t c2) > > so if you change that to > > cdef bint _mutability > > you probably only have to amend those 4 routines. Further attribute > access: > > sage/matrix$ grep \\._mutability *.pyx | wc > 20 69 1235 > > in matrix0.pyx these are just the lines in the functions above and the > remaining 9 lines are in a handful of special matrix implementations > and of the form > > if self._mutability._is_immutable: > > which aren't really adhering to protocol (accessing private > attributes! twice!), but are easily rewritten (to faster code anyway!) > > I thinks it's entirely doable to just make _mutability a straight > boolean flag. Does anyone have reasons why this is not a good idea? > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To post to this group, send email to sage-devel@googlegroups.com. To unsubscribe from this group, send email to sage-devel+unsubscr...@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel?hl=en.
