Confusing profiler results!

Hi!

I have used the gprof function to run a profile on minimad after
decoding a test mp3 file:

./minimad <test.mp3> test.pcm
gprof -b minimad.exe > profile_results.txt

The profiler results are attached at the end of this email. I have
compared the results to other profiler results in the mad-dev archives
and have found that my results differ from the others in that the
"output" function in minimad.c are at the top of the list when it comes
to "% time" and "cumulative seconds". Is it suppposed to be this way? (I
am running cygwin and using libmad-0.15.1b)

H.

Flat profile:

Each sample counts as 0.01 seconds.
  %   cumulative   self              self     total
 time   seconds   seconds    calls   s/call   s/call  name
 33.60      1.28     1.28     6752     0.00     0.00  synth_full
 17.06      1.93     0.65     6752     0.00     0.00  output
 13.65      2.45     0.52   486144     0.00     0.00  dct32
  9.19      2.80     0.35   523087     0.00     0.00  III_imdct_l
  7.87      3.10     0.30  2756891     0.00     0.00  mad_bit_read
  4.72      3.28     0.18    27008     0.00     0.00  III_huffdecode
  2.89      3.39     0.11    13490     0.00     0.00  III_stereo
  2.10      3.47     0.08    27008     0.00     0.00  III_aliasreduce
  2.10      3.55     0.08     6752     0.00     0.00  III_decode
  1.84      3.62     0.07  1140603     0.00     0.00  III_requantize
  1.31      3.67     0.05   523087     0.00     0.00  III_overlap
  1.05      3.71     0.04   432128     0.00     0.00  III_freqinver
  1.05      3.75     0.04                             __getreent
  0.79      3.78     0.03  1046174     0.00     0.00  fastsdct
  0.26      3.79     0.01    27008     0.00     0.00  III_scalefactors
  0.26      3.80     0.01     6753     0.00     0.00  mad_header_decode
  0.26      3.81     0.01                             putc
  0.00      3.81     0.00    33761     0.00     0.00  mad_bit_skip


--