Psyco tagging the same function multiple times

olsongt at verizon.net olsongt at verizon.net
Wed Jun 28 21:01:47 CEST 2006


skip at pobox.com wrote:
> I'm trying some stuff with Psyco and am confused about its apparent desire
> to compile the same function (or method) multiple times).  Here's an
> abstract from a recent run:
>
>     12:45:15.99  tag function: __main__.Watcher.processUpdate        %
>     12:45:24.95  tag function: __main__.Watcher.processUpdate        %
>     12:45:55.12  tag function: __main__.Watcher.processUpdate        %
>     12:46:12.29  tag function: __main__.Watcher.processUpdate        %
>     12:46:28.09  tag function: __main__.Watcher.processUpdate        %
>     12:47:05.32  tag function: __main__.Watcher.processUpdate        %
>     12:47:48.62  tag function: __main__.Watcher.processUpdate        %
>     12:48:09.63  tag function: __main__.Watcher.processUpdate        %
>     12:48:43.20  tag function: __main__.Watcher.processUpdate        %
>     12:49:58.80  tag function: __main__.Watcher.processUpdate        %
>     12:50:32.20  tag function: __main__.Watcher.processUpdate        %
>     12:51:07.91  tag function: __main__.Watcher.processUpdate        %
>     12:52:17.14  tag function: __main__.Watcher.processUpdate        %
>     12:52:33.67  tag function: __main__.Watcher.processUpdate        %
>
> What (if anything) does it mean that it tagged Watcher.processUpdate more
> than once?
>
> Thanks,
>
> Skip

I think that's what Armin means by a *specializing* compiler.  I
believe it will compile multiple versions based on the arguments going
into a block of code.




More information about the Python-list mailing list