Distributed Simulation Enviornment - Summary
This group is not part of the GNU Project.
distSim is a research project to construct a general purpose
distributed simulation environment. Attempting to answer the
question, "How to quickly and cost effectively, create an accurate
simulation if you aren't a programmer." Specifically, we have
biologists in mind for an early target audience. Please note, we are
in the early stages of design, and some of the project specific
terminology is not yet finalized.
The simulation consists of an environment, and models. Models exist
in the environment. The environment provides rules for how the models
co-exist, and a customizable ruleset. The environment will provide
for things like the gravitational constant, ambient air temperature,
etc.... Essentially, anything which affects all the models in a
geographic area.
Models are what need to get written for a new simulation. The
programmer will need to become familiar with the model API only. We
intend to ultimately provide bindings for several languages, but will
start with C and C++. The biologists will interact with a library of
already developed models when the software is mature.
Parallelism will principally be provided by assigning 'geographic'
boundaries, which overlap with network boundaries. That is to say,
within the simulation tile sets A and B are congruent, but exist on
separate machines. How these to sets interact will likely be the
topic of our first paper, but the biologists probably won't care.
Registration Date: Wed 21 Dec 2005 08:52:13 PM UTC
License: GNU General Public License v2 or later
Development Status: 2 - Pre-Alpha
posted by cshuller, Sun 20 Aug 2006 07:20:29 PM UTC
Yeay, the summer semester is over! That means I can get some work done. Currently my push is towards the User Interface, little messages are great for me, but make demonstrations no good.
The User interface I'm developing is in GTK. I expect to have it done by the end of the semester.
[...]
posted by cshuller, Tue 02 May 2006 03:38:14 PM UTC
Hey All,
It's finals week here at UCO. The project went through a successfull proof of concept, but we're now working on a mechanism by which to copy GObjects between MPI Tasks. I've been working on that for a couple of weeks, and am making good headway. The GObject code is found in ./gobject_move_test
posted by cshuller, Mon 06 Feb 2006 09:47:19 PM UTC
The meeting with Dr. Turner went well today. He approved of our work to date, and suggested we go with an interpreter to toss the models around with, rather than a complex NFS layout.
This alters the time table, but I am strongly inclined to bow to his wisdom in this matter.
[...]
posted by cshuller, Sun 05 Feb 2006 04:18:32 PM UTC
I spent all yesterday reviewing CORBA. It offers facilities many facilities which would be usefull for the modles implementation, but it 'externalises' objects by just shipping off their data.
In short, it would be usefull only in the case of the RM/DE -- MT communication. For that, we have already settled on SOAP.
[...]
- Browse open items
Submit a new item
Powered by Savane 3.14-8eb0.
Corresponding source code