Notes

Investigated ItemPerformance

I took a quick look at this by instrumenting parcels.osaf.framework.blocks.View#broadcast with hotshot.

Here are the results of starting up chandler and clicking around on dates in the mini calendar

>>> import hotshot, hotshot.stats
>>> stats = hotshot.stats.load("app.hotshot")
>>> stats.strip_dirs()
<pstats.Stats instance at 0x97850>
>>> stats.sort_stats('time','calls')
<pstats.Stats instance at 0x97850>
>>> stats.print_stats(30)
         171612 function calls (166184 primitive calls) in 5.740 CPU seconds

   Ordered by: internal time, call count

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
    11232    1.253    0.000    1.253    0.000 Item.py:143(__setattr__)
     9224    1.168    0.000    1.168    0.000 UUID.py:73(__eq__)
    11160    1.096    0.000    1.294    0.000 LinkedMap.py:356(iterkeys)
    11232    0.500    0.000    1.956    0.000 Item.py:486(getAttributeValue)
  5472/44    0.281    0.000    5.740    0.130 Views.py:44(broadcast1)
    11376    0.239    0.000    1.434    0.000 LinkedMap.py:107(_get)
    11232    0.160    0.000    2.116    0.000 Item.py:133(__getattr__)
     5428    0.158    0.000    0.158    0.000 Views.py:14(callMethod)
     5616    0.151    0.000    1.717    0.000 Views.py:101(<lambda>)
    11160    0.133    0.000    3.033    0.000 LinkedMap.py:351(__iter__)
    11232    0.109    0.000    0.109    0.000 Item.py:2266(_countAccess)
     5688    0.099    0.000    1.500    0.000 RefCollections.py:505(_getRef)
     5688    0.093    0.000    0.093    0.000 Item.py:1158(isNew)
    11160    0.077    0.000    1.371    0.000 RefCollections.py:842(iterkeys)
     5472    0.074    0.000    0.093    0.000 Values.py:355(_getRef)
     5688    0.037    0.000    1.308    0.000 LinkedMap.py:132(__getitem__)
     5472    0.035    0.000    0.035    0.000 LinkedMap.py:323(firstKey)
     5688    0.028    0.000    1.528    0.000 RefCollections.py:303(__getitem__)
    11376    0.026    0.000    0.026    0.000 UUID.py:60(__hash__)
     5472    0.011    0.000    0.011    0.000 RefCollections.py:49(_isUUID)
     5472    0.009    0.000    0.009    0.000 RefCollections.py:43(_isRefList)
       72    0.001    0.000    0.009    0.000 Views.py:83(<lambda>)
        0    0.000             0.000          profile:0(profiler)


<pstats.Stats instance at 0x97850>

The top 3 methods are Item.__setattr__, UUID.__eq__, and LinkedMap.iterkeys.

-- TedLeung - 22 Oct 2004