Conference azur::mcc

2285.0. "Survey: T1.2.4 MM activation time" by COOKIE::KITTELL (Richard - Architected Info Mgmt) Thu Feb 06 1992 12:06

T1.2.4 VMS

I don't have anything empirical enough to submit a QAR, but it seems like the
time from issuing a FCL directive to when the MM starts executing the 
directive entry point is much longer than with V1.1. I know this flies in
the face of having most of the T1.2.4 stuff, like the dictionary and parse
tables especially, faster than they were in V1.1 . The first few times I sat
here waiting to hit a debug breakpoint I figured it was simply a matter of
our systems being loaded more than when I was running V1.1. But the fact is,
once the MM is loaded all the operations are as fast as they ever were, and
that should be slower too if system loading was a factor.

It sure seems like something in the sequence of

        . parse the command using info from directory and parse tables
	. make the call through the function table
	. make the call through the access table
	. image activate the MM.EXE
	. call the init entry points
	. do MM-specific init and open MIR if any
	. call the directive entry point

is taking significantly longer. Has anyone else noticed this, or am I 
drinking too much coffee again? :-)

    I have noticed it as well and have voiced my "concern" over it.
    I see the base startup time (mange/enter -> prompt) having
    gone from abou 4-6 seconds to 10-12 seconds. With the debugger
    it is worse of course.
    
    But... the good news is, it is being worked on. Hopefully you
    will see some of that improvement in the next release (yes we
    are still planning on making an update available in house real
    soon).