Re: [Python-Dev] Python Specializing Compiler
Hi. [Armin Rigo] ...
Why it is related to Psyco: the current treatment of globals/builtins makes it hard for Psyco to statically tell what function we are calling when it sees e.g. "len(a)" in the code. We would at least need some help from the interpreter; at least hooks called when the module's globals() dictionary change. The above proposal might provide a more uniform solution.
FYI, a different proposal for opt. globals access by Jeremy Hylton. It seems, it would break fewer things ... don't know whether it can be as useful for Psyco: http://mail.python.org/pipermail/python-dev/2001-May/014995.html In any case I think Psyco will need notification support from the interpreter about dynamic changes to things that Psyco honestly assumes to be invariant in order to achieve performance. regards, Samuele Pedroni.
Samuele Pedroni
a different proposal for opt. globals access by Jeremy Hylton. It seems, it would break fewer things ...
I really like Jeremy's proposal. I've been having similar thoughts myself for quite a while. Greg Ewing, Computer Science Dept, +--------------------------------------+ University of Canterbury, | A citizen of NewZealandCorp, a | Christchurch, New Zealand | wholly-owned subsidiary of USA Inc. | greg@cosc.canterbury.ac.nz +--------------------------------------+
Samuele Pedroni
: a different proposal for opt. globals access by Jeremy Hylton. It seems, it would break fewer things ...
I really like Jeremy's proposal. I've been having similar thoughts myself for quite a while.
Greg Ewing
Ditto. Isn't this what I've been calling "low-hanging fruit" for ages? Apparently it's low but still out of reach. :-) --Guido van Rossum (home page: http://www.python.org/~guido/)
participants (3)
-
Greg Ewing
-
Guido van Rossum
-
Samuele Pedroni