#: 14600 S12/OS9/68000 (OSK)
    14-Mar-92  09:55:43
Sb: #14578-#MM/1 help
Fm: Steve Wegert 76703,4255
To: Kevin Darling 76703,4227 (X)

Kev,

This thing gets truely weird. I have eliminated all references to tsmon, any
version, and haven't had a lock up since. I'll be yanking the windowing stuff
today and booting straight to terminal, re-install the tsmon on the modem port
and see what happens.

I realize the amount of CPU time displayed on /term's shell is a reasonable
value in and of itself. The point I wanted to make was given the _nothing_ was
executing on the shell save for the echo's from the startup file, and all
activity for the 9 1/2 hours had been done on /t2's shell, the porportional
difference between the two appeared to be significant.

Another example from last night's session follows:

 Id PId Grp.Usr  Prior  MemSiz Sig S    CPU Time   Age Module & I/O
  2   0   0.0     128    0.50k   0 w        0.09 13:37 sysgo >>term
  3   4   0.0     128    6.75k   0 s        8.80 13:37 shell >>w1
  4   0   0.0     128    6.75k   0 w        0.09 13:37 shell >>w1
  5   2   0.0     128    6.75k   0 s     4:00.33 13:37 shell >>term
  6   0   0.0     128    6.75k   0 w        5.80 13:37 shell >>t2
  7   6   0.0     128   18.00k   0 *        0.12  0:00 procs >t2 >dd

As you can see, the shell on /term is crunching numbers big time, while the
shell on /t2 (where I was doing all my work) is fairly modest. Further, even
comparing the CPU time on the idle window (ID #3) which is brought up at boot
and just sitting there, to my active session on /t2 (ID #6) it would appear
that shells associated with the windowing stuff tend to get real busy.

Using Mark's nifty tool, Sysmon, which further breaks down CPU time into Total
CPU Time, System State, and User State, I see that only .05 of the time spent
on /term's shell is related to User activity. The rest being allocated in the
System State column.

I'll keep after this and let you know what I find.

Thanks for the help!

Steve




There is 1 Reply.