As for the speed: I would bother too much. real gains can be made when
you avoid using PUTS or PUTP to set a variable to the value it already
has: disk I/O is costly.

As mentioned in my first response: with one table name it is possible
to set & retrieve many variables at once, what is faster.  So, that's
what I'd use.  Furthermore: the table name was meant to separate
GLOBALV variables for different applications.  Issue GLOBALV GRPLIST
to see which table names already exist.

If you really want to know the difference: measuring isn't difficult,
I'll send you my BENCH EXEC.   You code two versions TEST1 and TEST2
EXEC and issue:
  BENCH EXEC TEST1
  BENCH (RUN NEXT) EXEC TEST2
and BENCH will display CPU costs in a small table.  Note  that the
TEST1 and TEST2 execs probably need to issue many GLOBALV calls to see
a noticeable difference: starting an EXEC is quite costly compared to
a simple GLOBALV call.


2008/1/18, Horlick, Michael <[EMAIL PROTECTED]>:
> I am a little confused in the difference between a SETS , for example and a 
> PUTS.
>
> My question is really about speed of retrieval. Is it faster searching one 
> group with many variables or have a group for each variable with one variable 
> per group?
>
> Thanks,
>
> Mike
>
-- 
Kris Buelens,
IBM Belgium, VM customer support

Reply via email to