Open MOdeL Experiment
Mathieu Leclaire
Romain Reuillon
Jonathan Passerat-Palmbach
The Complex system modelisation
These methods are time consuming
Naturally parallel algorithms permit to leverage parallelism
What does OpenMOLE do ?
Embed your model as a black box
A naturally parallel formalism to design experiments
Data parallelism
The parallelisation by message
The Grid
A workflow
The application
The console
Towards a multi-user web platform
Towards graphical editing
The Chromosome structuring
C++ model
2 days per simulation for 1600 simulations
= 8.5 years CPU time
The SimTRAP project
Netlogo model
5 min per simulation for 100000 simulations
= 1 years CPU time
The Simpop project
Scala model
2s min per simulation for
500 000 000 simulations
= 30 years CPU time
The Bioemergences project
C model - Portal access
Daily production
10000 hours / day
openmole.org
OpenMOLE Mailing List: list.openmole.org
Hands-on
openmole-market
Full examples available
Setup your environment
Download openmole: next.openmole.org
Don't close the tab you'll need it for the documentation ;)
~/.openmole
Stores your preferences
Maintains a table of file locations
Can grow rapidly in number of files
=> move it to a folder without quota restrictions
1st examplewww.doc.ic.ac.uk/~jpassera/openmole-tutorial-material
Pi Monte-Carlo calculation
1st example
Add variables declaration
Define exploration
Define environment
1st example
Remote environment?
PBS, Condor, Slurm, EGI, ...
Define Authentication
2nd examplewww.doc.ic.ac.uk/~jpassera/openmole-tutorial-material
Tree leaves classifier
Random Forest
Design
of
Experiments
2nd hands-on
Add variables declaration
Define exploration
Define environment
Collect results as CSV
Results visualisation?
Add another task
Runnning locally
Example: Python + Matplotlib scatter plot