[geclipse-dev] Trace View of the g-Eclipse project

[Thomas Köckerbauer <tkoeck@xxxxxxxxxx>]

Hello everyone

I just wanted to announce that the first version of the trace view plugin 
which is being developed as a part of the g-Eclipse project is in the 
g-Eclipse CVS repository in Karlsruhe now. (It is not yet in the SVN 
repository on eclipse.org, I guess this requires some legal process first). 
We do not have a built yet but I think this should be the case soon.

The plugins are:
- eu.geclipse.traceview
  This is the main plugin, which provides the physical clock
  and the lamport clock visualisation of the trace data. It
  also contains some basic markers (these can be used by
  plugins to mark regions of interest).

  Physical graph:
  http://www.gup.jku.at/~tkoeck/traceview_physical.png

  Lamport graph:
  http://www.gup.jku.at/~tkoeck/traceview_lamport.png

  A marker that shows the cause and effect relations on a
  selected event:
  http://www.gup.jku.at/~tkoeck/traceview_marker.png

  The plugin requires some g-Eclipse core plugins from the
  SVN repository on eclipse.org.

- eu.geclipse.traceview.debug
  This plugin provides the debugging functionality for MPI
  programs started on a remote grid site using the glogin tool.
  This is the part where we think it would make most
  sense to reuse parts of PTP or to add the glogin
  connection type (maybe through some abstraction) for
  remote debugging to the PTP tools.
  This plugin has a dependency to CDT.

- eu.geclipse.traceview.nope
  A trace reader for the NOPE trace format.

  Here is an example project+trace:
  http://www.gup.jku.at/~tkoeck/poisson.tar.gz

- eu.geclipse.traceview.otf
  A very basic trace reader for the OTF file format.
  It can only open a few OTF files yet.

- eu.geclipse.traceview.statistics
  An example plugin for the extension of the trace view by
  other visualisation types, this plugin provides extension
  points for statistic data providers that generate statistical
  data out of the trace and for visualisation types that can
  visualise the different types of data.

  Screenshot:
  http://www.gup.jku.at/~tkoeck/traceview_statistics.png

  This plugin has a dependency to BIRT


If you have questions feel free to ask. If you are at the parallel tools 
workshop in Stuttgart we can also talk about it there if you like.


Regards,
Thomas