Can you tell me how much of a performance difference might exist, if any, between these two alternative RHDCUXIT definitions?
When you define a numbered exit in RHDCUXIT with EP=, you specify the entry point and hard link the exit with RHDCUXIT.
The exit is loaded at startup when RHDCUXIT and other nucleus modules are loaded.
When you instead define with NAME= the exit program must be linked as a loadmodule into a CDMSLIB loadlib and must be defined as a PROGRAM in sysgen.
The first time the exit is called it is loaded.
If you have properly sized program pools (in this case it would be the XA Reentrant Pool) then no program would ever need to be loaded more than once and would stay loaded.
There should be no noticeable performance difference due to having to do the one and only load of the exit module.
Also, using NAME= gives you more flexibility if the exit ever needs to change. You would only have to reassemble/link it and then DCMT V PRO xxxxxx NC instead of having to relink RHDCUXIT and a Nucleus Reload or CV recycle.
SYSTEM OPERATIONS GUIDE USER EXITS, NUMBERED EXITS
Release: IDADSO00100-18.5-ADS-for CA-IDMS