Profiler, profiling DLLs, trace.log

[Bjoern]

Lutger schrieb:
> Bjoern wrote:
> 
> ...
>> I guess the numeric data at the bottom part are relevant. And each of
>> them is calulated based on PART I.
>>
>> How ? NONSENSE ?
> 
> iirc, the timings in part one are in ticks, not microseconds, while the
> timings in the summary are calculated to be in microseconds. This is a bit
> confusing. In the table of the second part the frequency is stated which
> can be used to calculate from ticks to microseconds. 
>  
>> Last Q for the moment. What is PART I good for ? Or in other words, Can
>> I use this information somehow ?
> 
> Call graphs. For each function, it tells you by which functions it is called
> and which functions it calls, along with timing information. This can give
> useful hints when looking for performance bottlenecks.
> 
> 
Thanks Lutger.
pTrace actually just supports the summary table, right ?

Call graphs.
I wonder how a graphical representation should look like.
Perheps:

called from

\ /
  F
/ \
     /0.012ms/
calls


Ideas ?
Bjoern