The following image is an example of histogram whose data came from MILC-benchmarks drawed by Octave.

This image shows a clock_cycle histogram of message-passing and mat-vec multiply in the dslash_fn and dslash_fn_special function. These commands recorded below.

------------------------------------------------------------------------
qcdhome> octave
GNU Octave, version 2.0.16 (i686-pc-linux-gnu).
Copyright (C) 1996, 1997, 1998, 1999, 2000 John W. Eaton.
This is free software with ABSOLUTELY NO WARRANTY.
For details, type `warranty'.

octave:1> load mpassing.oct
octave:2> load masmv4.oct
octave:3> load msmvs4.oct
octave:4> whos

*** local user variables:

prot  type               rows   cols  name
====  ====               ====   ====  ====
 wd   matrix             5728      1  masmv4
 wd   matrix            49120      4  mpassing
 wd   matrix             5728      1  msmvs4

octave:5> mhist = histogram(mpassing(:,1), 1260);
octave:6> chist = histogram([masmv4', msmvs4']', 1260);
octave:7> gset nokey
octave:8> gset title "12^4 lattice 4np(Athlon 1.2GHz) by mpich1.2.1..7"
octave:9> gset xlabel "clock cycle (bin=1260:0.000001sec)"
octave:10> gset ylabel "number of events"
octave:11> gplot mhist with impulses, chist with impulses
------------------------------------------------------------------------

A generated mpassing.oct has 4 columns, clock cycle, data size, dist, and source. If you pick up longlinks or fatlinks in the data, you use 2nd column data.

You can get the trace data of this example in here.