Boost logo

Boost :

From: Ullrich Koethe (koethe_at_[hidden])
Date: 2000-10-16 05:43:07


Hi Prabhu,

> Ullrich> Yes, but the smaller the granularity of the wrapped
> Ullrich> functionality, the more the complexity of this
> Ullrich> abstraction layer will hurt performance. A ridiculous
> Ullrich> example: imagine your WWW browser would open an ILU
> Ullrich> connection for every single pixel of each image - this
> Ullrich> would be way too slow. So you will need a thinner
> Ullrich> abstraction layer for finer granularity (but probably
> Ullrich> py_cpp is still to slow to routinely access single
> Ullrich> pixels).
>
> Not if the ILU was used to compile a new library that deals
> with the python calls directly. What I really meant was that one
> generates the library just as py_cpp does now but with a layer of
> abstraction that is used _during_ the conversion of the library and
> _not_ during runtime. This may have problems of its own and will
> certainly add complications.
>

It is indeed an interesting idea to push these things towards compile
time. Currently, there is hot research (e.g. in the areas of 'active
libraries' and 'partial evaluation') to find out how one can do context
specific, customized optimizations at compile time, in order to minimize
the amount of work left at run-time. However, as far as I can tell, this
has not matured enough for mainstream applicability.

In the context of Python, it wouldn't help me too much, because I need
Python to gain *run-time flexibility* that, by its very nature, is not
susceptible to compile-time optimizations. (However, it could profit
from just-in-time compilation that is also investigated in the context
mentioned above.)

Ulli

-- 
 ________________________________________________________________
|                                                                |
| Ullrich Koethe  Universität Hamburg / University of Hamburg    |
|                 FB Informatik / Dept. of Computer Science      |
|                 AB Kognitive Systeme / Cognitive Systems Group |
|                                                                |
| Phone: +49 (0)40 42883-2573                Vogt-Koelln-Str. 30 |
| Fax:   +49 (0)40 42883-2572                D - 22527 Hamburg   |
| Email: u.koethe_at_[hidden]               Germany             |
|        koethe_at_[hidden]                        |
| WWW:   http://kogs-www.informatik.uni-hamburg.de/~koethe/      |
|________________________________________________________________|

Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk