From: Andreas Kupries <andreask@ac...> - 2010-04-05 21:26:29
Guilherme Polo wrote:
> 2010/4/5 Matthew M. Burke <mmburke@...>:
>> Guilherme Polo wrote:
>>> problems you and Donal are mentioning. Maybe for Tcl it might not be
>>> worth it, but it is hard (to me) to know for sure if a JIT compiler
>>> won't help Tcl at all.
>> I've not given the matter much thought, but Mike Pall has had a lot of
>> success with a JIT for Lua (http://luajit.org/). I would think
>> you should be able to get similar results for Tcl.
> LuaJIT, especially the second version, does seem to bring a lot of
> improvement to Lua with a JIT compiler but it also has implemented a
> Lua interpreter in assembly and shows great speedups with it too. It
> does way more than what I've mentioned earlier here, and I don't think
> I could propose something like that and be realistic with the
IMHO anything which is interesting and you believe cannot be done within our
timeframe (of one summer) can go into a section of the application called
'Future Directions'. IOW I believe that you should at least mention these
things, even if you consider them out of scope of what you are trying to do.
Andreas Kupries <andreas@...>
Developer @ <http://www.activestate.com/>