I am using Regina on Unix (Sun/Solaris) and we have recently run into performance problems caused by the amount of data we require our script to process.
It looks as if it gets worse as the amount of variables created increases. This does not seem to be a linear process.
I have searched a little bit and found somebody on the Web mentioning this..
Regarding the performance of Rexx, participant Ian Collier pointed out that the otherwise high performance Regina interpreter suffers from its fixed hashtable size for the phonecode problem, because the default size of 256 is too small. Increasing this value to 8192 (which requires recompiling Regina) reduced the run time of Collier's Rexx program from 53 seconds down to 12.
Does anybody know where to find isntructions as to how Regina can be recompiled and what is required to do that? Where do I need to adjuts the hastable size?!
Thanks a lot for your help and advice.