Re: 41-MCODE: SOLVE & INTEG - 4k ROM Message #6 Posted by 聲gel Martin on 17 Apr 2012, 2:45 p.m., in response to message #5 by Massimo Gnerucci (Italy)
Guys, guys, you're going to make me blush! Thanks for your kind words but it's not such a big deal, really. Truth is as much as I enjoy writing NEW MCODE functions for the Coconut, I also dig going into this archaeological kind of hunting - it keeps my aging brain active :-)
This particular example was fun to do. I managed to squeeze about 60 bytes in HP's code just by removing a lot of SETHEX instructions, replacing them with the appropriate port_dependent call (you know, starting with 375 as opposed to 379 - within the same quad.) Needless to say those 60 bytes were used up in this module with other stuff.
I can say I sort of understand about 85% of the code, so there's another 15% chance that it flunks :-) I tested it with several examples though, like the nested Bessel J zeros, a rough acid test for sure - and so far it's holding up nicely.
Cheers,
'AM.
|