Unit HIGH PERFORMANCE COMPUTING
- Course
- Informatics
- Study-unit Code
- 55961406
- Curriculum
- In all curricula
- Teacher
- Osvaldo Gervasi
- Teachers
-
- Osvaldo Gervasi
- Hours
- 78 ore - Osvaldo Gervasi
- CFU
- 9
- Course Regulation
- Coorte 2023
- Offered
- 2023/24
- Learning activities
- Caratterizzante
- Area
- Discipline informatiche
- Academic discipline
- ING-INF/05
- Type of study-unit
- Obbligatorio (Required)
- Type of learning activities
- Attività formativa monodisciplinare
- Language of instruction
- English
- Contents
- Introduction to High Performance Computing (HPC) and High Throughput Computing (HTC); Cluster implementation and administration. Construction of scalable, high-reliability services. The Condor Project for High Throughput Computing. Cloud Systems Architecture; Process Management and Balancing in Cluster and Cloud environments; GPU Computing; OpenCL. Cloud Computing. Virtualisation, Docker, Kubernetes Systems.
- Reference texts
- J. Garrison and K. Nova, Cloud Native Infrastructure-Patterns for Scalable Infrastructure and Applications in a Dynamic Environment, O’Reilly and Associates, 2018, ISBN: 978-1-492-03969-3
Teaching material provided by the lecturer with bibliographical references - Educational objectives
- The main objective of the course is to introduce the student to the study of emerging systems, which are among other things the basis for the realisation of modern data centres and the release of high reliability environments, as well as to present high performance systems, which are fundamental for the advancement of knowledge and science, and for delivering modern services, including those in the mobile environment.The main knowledge acquired will be:Design and implementation of Cluster systems: main problems and methods of installation and configuration High reliability services and high performance systems The Condor environment for the realisation of High Throughput Clusters. GPGPU Computing and GPU OpenCL programming: to increase the performance of many applications. Cloud Computing: design and implementation of hybrid clouds. Virtualisation environments, Docker and Kuberntes systems.
- Prerequisites
- No prerequisites are required to the student to attend the course; however the knowledge related to designing and implementing Internet Services it is expected to be known by all students.A strong knowledge of open source operating systems like Linux facilitates the acquisition of the concepts presented during classes. Furthermore, examples and use cases will be presented, aimed at facilitating the learning of the concepts covered during classes.
- Teaching methods
- Lectures using multimedia tools (slides, video, etc). Exercise in the computer room.
- Learning verification modality
- The exam consists of a written test on the Libreeol platform (https://libreeol.org) followed by an oral exam of approximately 30 minutes in which the student must illustrate a project implemented and agreed upon with the lecturer, demonstrating mastery of the required concepts, adequate language property and the ability to organise the concepts to be expounded in an appropriate manner.
- Extended program
- The course introduces students to modern technologies that enable the implementation of data centres and the delivery of applications in high-reliability and load-balancing environments. In particular, the following topics will be covered: Introduction to High Performance Computing (HPC) and High Throughput Computing (HTC); Beowulf clusters; Implementation and administration of Clusters: Building scalable services; Implementation of high reliability services in a Cluster environment: Heartbeat, Corosync, PaceMaker. Condor: Installation, implementation and design of a High Throughput Cluster. GPGPU Computing. OpenCL programming language. Cloud Computing: presentation and implementation of a Hybrid Cloud. Virtualisation systems; Docker; Kubernetes systems.
- Obiettivi Agenda 2030 per lo sviluppo sostenibile
- This course contributes to the realisation of the UN goals of the 2030 Agenda for Sustainable Development