From: Joel VanderWerf <vjoel@...>
Date: 2002-10-29T04:12:02+09:00
Subject: Re: interpreting the second profiling column of Ruby profiler

Olonichev Sergei wrote:
> Hello,
> 
> I doubt in interpreting the second profiling column of Ruby profiler.
> Here you are a small portion of the profiler's output:
> 
>   %   cumulative   self              self     total
>  time   seconds   seconds    calls  ms/call  ms/call  name
>  12.63   115.42    115.42   268703     0.43     0.63  Array#hash
>   7.48   183.72     68.31   160340     0.43     0.64  Array#eql?
>   6.06   239.10     55.38   161270     0.34     1.56  Hash#[]
>   5.99   293.87     54.77   537406     0.10     0.10  Kernel.hash
>   4.29   333.06     39.19    14370     2.73    10.98  FST#del_transition
> 
> [...]
> 
>   0.00   905.65      0.00        1     0.00 913743.00  #toplevel
>   0.00   905.65      0.00        5     0.00     0.00  Module#public
>   0.00   905.65      0.00        6     0.00     0.00  Hash#clear
>   0.00   905.65      0.00        1     0.00   460.00  Chains2MINFST#finish
>   0.00   905.65      0.00        5     0.00     0.00  Hash#new
> 
> In "Ruby Developer's Guide", it is written:
> "... The second column is the cumulative for the methods up to and including
> the current one. ..."
> 
> But why the numbers below in the second column are so big?

It's cumulative as you read down the column, adding "self seconds" each 
time. So:

  183.72 = 115.42 + 68.31  (approx.)

and so on.