Program

Parallel and Distributed Computing Essentials
Time	Description
09:00 - 11:00	Parallel Architectures and HPC Basic Concepts
11:00 - 11:30	Part I: Cluster Services and Installation Procedures
11:30 - 12:00	Part II: NETBOOT Environment and Troubleshooting
12:00 - 12:30	Part III: Cluster Management Tools and Security
12:30 - 13:00	Part IV: Hands-on_Laboratory
13:00 - 14:00	Lunch
14:00 - 16:30	Laboratory Part I: Installing a cluster and configuring cluster services
16:30 - 19:00	Laboratory Part II: Parallel environment and resource management system
19:00 - 20:00	Dinner
20:00 - 22:00	Laboratory Part III: Serial and parallel job submission with Torque/Maui + openMPI

Parallel Architectures and HPC Basic Concepts

09:00 - 11:00
Speaker: Gilberto Diaz

Through this presentation attends will have several lectures on the basis of HPC and distributed computing.

Part I: Cluster Services and Installation Procedures

11:00 - 11:30
Speaker: Moreno Baricevic

A brief introduction to HPC clusters, covering both the HW and the SW infrastructure, and an overview of some installation procedures for clusters and the services involved.

Part II: NETBOOT Environment and Troubleshooting

11:30 - 12:00
Speaker: Moreno Baricevic

Configuration, setup and troubleshooting of the network booting environment for remote, distributed and unattended deployment of computers. Overview of PXE, DHCP, TFTP, NFS, package repositories, Kickstart/Anaconda, system logging.

Part III: Cluster Management Tools and Security

12:00 - 12:30
Speaker: Moreno Baricevic

Overview of some management tools and protocols needed for the administration and the monitoring of the cluster status and the services involved; brief notes on security.

Part IV: Hands-on_Laboratory

12:30 - 13:00
Speaker: Moreno Baricevic

A brief introduction to the today's laboratory session about cluster installation and configuration.

Laboratory Part I: Installing a cluster and configuring cluster services

14:00 - 16:30
Speaker: Moreno Baricevic

The aim of this exercise is to install and configure the services needed to setup an unattended installation of a cluster. Each group of students is expected to configure the masternode, install missing packages and remove unwanted ones, configure the required services, and finally install one computing nodes. The same cluster will be used for the second part of the lab.

Exercises: http://edu.escience-lab.org/SCCamp2011/ClusterInstallationPart1

Laboratory Part II: Parallel environment and resource management system

16:30 - 19:00
Speaker: Moreno Baricevic

The aim of this exercise is to install and configure the batch system and the scheduler in order to setup a queue system to a parallel environment. The cluster set up during the first part of the lab is now going to be fully configured as a real-world HPC cluster. Each group of students is expected to download, compile and install the components of the queue system, and setup such services. It's also asked to create a test account for a user and setup NFS in order to export some filesystems. For administration tasks, the students are supposed to setup a passwordless environment and install a parallel shell.

Exercises: http://edu.escience-lab.org/SCCamp2011/ClusterInstallationPart2

Laboratory Part III: Serial and parallel job submission with Torque/Maui + openMPI

20:00 - 22:00
Speaker: Moreno Baricevic

The exercise will be considered successfully completed when the "user" will be able to submit a batch job and run an mpi parallel test on both the computing node and the masternode installed during the first part of the lab. Installing openmpi has to be considered part of the exercise.

Exercises: http://edu.escience-lab.org/SCCamp2011/ClusterInstallationPart2