[alexannika@users.sourceforge.net: Treemap Visualization for profiling output]

["H . J . Lu" <hjl at lucon dot org>]

I think it belongs here, not bug-gnu-utils@gnu.org. That brings out
another question. Where should a user report a binutils bug? Right
now, we tell people to

Report bugs to bug-binutils@gnu.org

But bug-binutils@gnu.org is a catch-all mailing list, which covers
more than just binutils. I am not sure how many binutils developers
are even on bug-binutils@gnu.org. Shouldn't bug-binutils@gnu.org be a
separate bug report mailing list for binutils? In fact, I personally
don't mind the REAL bug reports on binutils@sourceware.cygnus.com.


H.J.
----
----- Forwarded message from Alexander Rawass <alexannika@users.sourceforge.net> -----

Delivered-To: hjl@lucon.org
From: Alexander Rawass <alexannika@users.sourceforge.net>
Subject: Treemap Visualization for profiling output
Date: Mon, 30 Jul 2001 05:23:53 +0200
X-Mailer: KMail [version 1.2]
To: hjl@lucon.org, bug-gnu-utils@gnu.org

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I got your email adress by 'gprof --help', and I'd like to talk to you about 
using my Treemap-Widget to display output from programs like gprof in a 'new 
way'.

I am the author of QTreeMap, a general Qt/KDE-Widget to display general 
hierarchies (that can be associated with a size) in from of a Treemap.

Treemaps are a 'new' way of displaying a hierarchical tree in a graphics area.
Each node in the tree is drawn as a rectangle, with its subnodes draws inside 
it, all drawn proportional to their size - so, large nodes get large areas on 
the screen and small ones small areas.
You can see the _whole_ hierarchy of the tree at _one_ glance as a 800x600 
painting.

Screenshots of my /usr/X11R6 filetree seen as a treemap:
http://kdirstat.sourceforge.net/kdirstat/treemaps/screenshots/sqr_bump_usr_x11r6.png
http://kdirstat.sourceforge.net/kdirstat/treemaps/screenshots/flat_usr_x11r6.png
http://kdirstat.sourceforge.net/kdirstat/treemaps/screenshots/dist_cushion_mono_usr_x11r6.png

The output of gprof can be seen as hierarchical data, where each element is a 
block of the program that has been accessed n times or for t seconds.
If profiling output would be displayed as a treemap, you could see at one 
glance which code-blocks use the most time - they are the rectangles with the 
largest size/area.

QTreeMap is currently part of KDirStat, a disk-usage utility.
QTreeMap is used to display the size of a directory and all its subdirs and 
their files at one glance - you can easliy pick out which are the largest 
files or dirs on your harddisk.

I am not familiar with gprof or profiling or the format of the profiling 
data, but I really think that a Treemap could display this profing data 
better than the normal ascii output.

So, I'm searching for people to help me, to write a new subclass of QTreeMap 
that can work on a tree with profiling data.

If've already tried to contact the author of KProf 
(fpillet@users.sourceforge.net), a profiling data viewer with a traditional 
treeview, but he hasn't asnwered yet.

Do you know of other profile viewers for X,Gtk/Gnome or Qt that I could try 
to modify?
Or could I even use gprof itself as the basics for this viewer application?
In any way, I need help from those who are familiar with prof. output.


Links:

KDirStat         http://kdirstat.sourceforge.net/kdirstat/
QTreeMap         http://kdirstat.sourceforge.net/kdirstat/treemaps/index.html
KDirStat CVS: 	http://sourceforge.net/cvs/?group_id=30290
SequoiaView, Treemap Viewer for Win32   	http://www.win.tue.nl/sequoiaview/
Home of the inventor of Treemaps:  	http://www.cs.umd.edu/hcil/treemaps/


I hope to hear from you



Alex


- -- 
                   Alexander Rawass
          Email: alexannika@users.sourceforge.net
     Project Homepage: http://tuxfleet.sourceforge.net
...but some day you'll be a STAR in somebody else's SKY...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE7ZNNRwQI8aoth0hwRAsxEAJ4km+gEkvIaUCB9XK1jvK9Nm4uJPgCgh6x8
sG3gLx1iVeD/mBlqavhFkPA=
=65ns
-----END PGP SIGNATURE-----

----- End forwarded message -----