I've noticed that the run times reported by the "stats" command are
consistently completely wrong in Linux. For example, I compiled the trunk
kernel (r13874) with optimizations, and ran the 15 puzzle agent for about 1
minute. I disabled all output and removed the monitor rule, so no time was
spent printing to the terminal. The output of stats after this is:
Soar 9.3.2 on puff at Mon Jun 24 20:30:39 2013
98 productions (73 default, 13 user, 12 chunks)
+ 0 justifications
| Computed
Phases: Input Propose Decide Apply Output | Totals
========================================================|===========
Kernel: 0.001 0.001 0.009 6.006 0.001 | 6.018
========================================================|===========
Input fn: 0.001 | 0.001
========================================================|===========
Outpt fn: 0.000 | 0.000
========================================================|===========
Callbcks: 0.000 0.000 0.000 0.000 0.002 | 0.002
========================================================|===========
Computed------------------------------------------------+-----------
Totals: 0.002 0.001 0.009 6.006 0.003 | 6.021
Values from single timers:
Kernel CPU Time: 6.023 sec.
Total CPU Time: 6.028 sec.
8521 decisions (0.707 msec/decision)
25634 elaboration cycles (3.008 ec's per dc, 0.235 msec/ec)
25634 inner elaboration cycles
8532 p-elaboration cycles (1.001 pe's per dc, 0.706 msec/pe)
148508 production firings (5.793 pf's per ec, 0.041 msec/pf)
282374 wme changes (141293 additions, 141081 removals)
WM size: 212 current, 210.842 mean, 371 maximum
Running the same test for 30 seconds and 120 seconds, I noticed that the times
reported are always approximately 1/10 of the real time. So I'm guessing that
somewhere in the code there's a division that's missing a 0. If that's the
case, the fix should be simple.
I've confirmed that the times reported in Windows are correct, but I haven't
tried on Mac.
Original issue reported on code.google.com by joseph...@gmail.com on 25 Jun 2013 at 12:57
Original issue reported on code.google.com by
joseph...@gmail.com
on 25 Jun 2013 at 12:57