unexpected results from tcllib profiler?
d-list-9_at_jendaveallen.com
Date: 03/26/05
- Next message: Ian Bell: "Re: NDA agreements"
- Previous message: Bryan Oakley: "Re: Simple ? about wm withdraw"
- Next in thread: d-list-9_at_jendaveallen.com: "Re: unexpected results from tcllib profiler?"
- Reply: d-list-9_at_jendaveallen.com: "Re: unexpected results from tcllib profiler?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 26 Mar 2005 07:28:28 -0800
Folks,
I am a little confused about the tcllib profiler package. I am using
the 8.4.9 distribution on Windows XP, which includes profiler 0.2.2.
Here is my elaborate test program.
package require profiler
::profiler::init
proc c1 { a } { for {set i 0} {$i<2000} {incr i} {;} }
proc c2 { a } { for {set i 0} {$i<2000} {incr i} {;} }
proc parent { } { for {set i 0} {$i<2000} {incr i} { c1 $i ; c2 $i }
}
parent
::profiler::dump ::parent
The output is:
::parent {callCount 1 callerDist {GLOBAL 1} compileTime 2816000
totalRuntime 2816000 averageRuntime 0 stddevRuntime 0
covpercentRuntime 0 descendantTime 0 averageDescendantTime 0
descendants {}}
The descendants data I would expect is like {c1 2000 c2 2000}.
The time reporting also seems wrong; how can all the time be
reported as compile time with zero descendant time? All the
time is spent in the children.
Am I missing something, or is this output completely wrong?
- Dave Allen: d-list-9@jendaveallen.com
- Next message: Ian Bell: "Re: NDA agreements"
- Previous message: Bryan Oakley: "Re: Simple ? about wm withdraw"
- Next in thread: d-list-9_at_jendaveallen.com: "Re: unexpected results from tcllib profiler?"
- Reply: d-list-9_at_jendaveallen.com: "Re: unexpected results from tcllib profiler?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]