Training session on
Scilab Parallel Computing with ProActive Parallel Scientific Toolbox
Date: 25-26 March 2015
Venue: Trity Technologies Training Center, Selangor
“Distribute your Scilab tasks to other workers with minimum programming involved!”
With ProActive Parallel Scientific Toolbox, you can accelerate your scientific and statistical computations. As modern scientific and engineering problems grow in complexity, the computation time and memory requirement increase and parallel computing becomes a necessity. ProActive integrates with de-facto standards in scientific and engineering environments such as Matlab, Scilab and R.
Directly from within these familiar environments, it provides users with the capacity to parallelize executions and manage data transfers on other Desktop machines, Clusters, Grids and Clouds. A single tool for accelerating all your scientific languages.
This two-day course provides a jumpstart on how to setup your own machine for parallel processing with Scilab. This training would focus on how to send parallel tasks from Scilab to speed up the computation time.
Who Must Attend
Researchers who are interested in parallel computation and wish to learn how to setup and run the computations faster with existing resources. This training would use Scilab as the computation software.
Basic Scilab, Matlab or R knowledge
What you will learn
How to run your computations faster with your existing resources
· How to setup a single infrastructure for your Matlab, Scilab, R and general purpose computations
· How to perform computing automation with workflows
· How to connect and interface your computations with other environments.
Day 1 :
Introduction to ProActive Parallel Scientific Toolbox
Objective : Give an overview of the ProActive technology and course themes.
· Grid and parallel computing
· Multiusers platforms and scheduling
· Workflows, tasks, and data management
Objective : Setup your ProActive Parallel Scientific Toolbox environment and become familiar with its main features. This environment will be used all along the course.
· Installation and setup of a ProActive infrastructure on your machine
· Installation of ProActive Parallel Scientific Toolbox into Scilab
· Execution of your first parallel computation right from your Scilab Console
Working with the ProActive interfaces
Objective: Learn how to use the best of your ProActive infrastructure to reduce execution time of your computations, while optimizing the use of your computing resources.
· Usage of PAsolve command
· Interactions with the ProActive Scheduler and Resource Manager from Scilab
· Understand the scheduling policies
· Add, remove and manage extra computing resources
· Setup a grid with multiple machines
· Understand and use the various ProActive web portals
Objective: Understand how your computations are parallelized to get most benefits from it.
· Start computations with asynchrony
· Use futures and synchronization on results to optimize parallelization
· Use disconnected mode to let jobs running over the weekend
Day 2 :
Working with tasks and functions
Objective: Use tasks and functions in your Scilab program to get a perfect control on parallelization and data transfers
· Parallelization with parametric sweep
· Use multiple ways to create tasks depending on the use case
· Compose functions
· Understand closures
Working with data
Objective: Learn how to use common data structures (including input and output files) with the ProActive toolbox commands for parallelization and data transfers.
· Use the PAsolve function
· Understand how variables, vectors and matrix are split for task parallelization
· Transfer data from input files and write results in output files
· Work on local and remote data with seamless transfers
Working with workflows
Objective: Build workflows to automate and orchestrate various kind of tasks (Scilab, R or any script) and integrate them in your existing environments.
· Design workflows with ProActive Studio
· Create various kind of tasks, including Scilab, R and Python
· Use parametric workflows
· Integrate them in your environment with ProActive APIs
Actual trainee case studies
Objective: General discussion with eventual hands-on part where each trainee describes its use case, so instructor explains how ProActive can help him.
The above information is brought to you by Trity Technologies Sdn Bhd in collaboration with Activeeon, developer of the ProActive Parallel Scientific Toolbox.
Trity Technologies Sdn Bhd 874125-T
26-3 Jalan Puteri 2/4
Visit us at www.tritytech.com
2000 Rt. des Lucioles
Les Algorithmes - Pythagore B - 2ème étage
06560 Sophia Antipolis
Phone: +33 (0)9 88 777 660
Fax: +33 (0) 9 88 777 661
Visit us at www.activeeon.com