[Python-Dev] Re: a serious threat to 2.3's speed?

[Alex Martelli]

On Thursday 11 December 2003 04:09 pm, Skip Montanaro wrote:
>     >> Another option for a masters thesis Brett - speeding up function
>     >> calls :)
>
>     Fredrik> that would be most excellent.
>
> Indeed.  We should make sure we're comparing apples to apples when looking
> at Jim's benchmark figures though.  It's not clear to me that his current
> system supports all the different types of callables which CPython
> supports. That alone might streamline his function call code a lot.

How so?  Isn't one "indirecting" through the tp_call slot anyway?  Supporting
two or two thousand types of callables should make function calls just about
the same cost -- or am I missing something...?


Alex