A faster Python?, Python compiler, Dylan,...
Michael Hudson
mwh at python.net
Wed Apr 3 12:07:48 EST 2002
Jan Kybic <kybic at ieee.org> writes:
> > > It would be so much easier if I could just compile parts of the
> > ^^^^
> > You make it sound so simple...
>
> Well, translating something like:
>
>
> def fib(a):
> if a <= 2: return 1
> else: return fib(a-2) + fib(a-1)
>
> into:
>
> int fib(int a) {
> if(a <= 2) return 1;
> else return fib(a-2) + fib(a-1); }
>
> does indeed seem to be simple, once you know that 'a' is an integer.
Just picking, but the C has different semantics wrt. overflow...
> > Have you looked at psyco?
>
> Yes, very impressive once it does what it promises.
I'm not sure what promises you mean, but it gets a 4x speedup on
pystone today.
> It is more ambitious than what I had in mind in that it does not use
> any assistance (type information) from the programmer.
Yes.
> > Well, there was boundless (it made me exceed my quota on the mail
> > server I was using ot the time) discussion on the types-sig on
> > more-or-less this sort of topic a couple of years back, although
> > perhaps more of the focus was on guaranteeing correctness than
> > optimization. You could read that, but you don't want to.
>
> Yes, I remember.
Ah.
> I did not quite buy the correctness argument, runtime checking is
> fine with me, but for a compiler it might be very useful, IMO.
I was more interested in the optimization too..
> > It's possible, perhaps even desirable, that psyco could be extended in
> > this fashion. Got a spare man-month or so? I haven't.
>
> No, neither have I, sorry. :-( Maybe later.
Indeedy. Maybe after 2.2.1 is done...
Cheers,
M.
More information about the Python-list
mailing list