On Mon, Jan 30, 2012 at 8:16 AM, William Stein <wst...@gmail.com> wrote: > > On Jan 30, 2012 8:09 AM, "Nathann Cohen" <nathann.co...@gmail.com> wrote: >> >> Helloooooooo !! >> >> >> > If you ever change a filebfrom Python to Cython for no clear good >> > reason, I >> > will be very annoyed. >> >> Well, the point's always to be able to rewrite code more efficiently :-p >> >> >> > One should always use Python instead of Cython unless >> > there is a clear compelling reason not to. >> >> Really ? O_o > > Yes. > > I would even add this to the review checklist. > >> >> > - takes a long time to compile >> >> Come ooooooon ! Half a second when you rebuild Sage !!! How bad can that >> be ? >> > > It takes nearly an hour on one cpu to build the sage library now right? > If you change all .py to .pyx, it might take 2 hours. That .5 seconds you > site can be longer for bigger files and adds up.
More precisely, right now on sage.math doing "unset MAKE; sage -ba" takes 46 minutes: time ./sage -ba ... real 46m22.127s user 43m40.140s sys 2m4.770s If you made it so that all Sage .py files were .pyx files, that would probably increase to about 2 hours, at least. For people using computers with a "limited" number of cores (e.g., 2), doing "sage -ba" is quite painful. I'm sure there is a ton of pure Python code that is in .pyx files for no good reason, which is already slowing things down. Let's not make it worse on purpose. -- William -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
