HPC and Grid Concepts



Similar documents
Cluster, Grid, Cloud Concepts

An approach to grid scheduling by using Condor-G Matchmaking mechanism

The Lattice Project: A Multi-Model Grid Computing System. Center for Bioinformatics and Computational Biology University of Maryland

Concepts and Architecture of the Grid. Summary of Grid 2, Chapter 4

Introduction to grid technologies, parallel and cloud computing. Alaa Osama Allam Saida Saad Mohamed Mohamed Ibrahim Gaber

Grid Scheduling Architectures with Globus GridWay and Sun Grid Engine

Grid Computing vs Cloud

Storage Virtualization from clusters to grid

Distributed Systems and Recent Innovations: Challenges and Benefits

Web Service Based Data Management for Grid Applications

The Data Grid: Towards an Architecture for Distributed Management and Analysis of Large Scientific Datasets

GRID COMPUTING Techniques and Applications BARRY WILKINSON

OBJECTIVE. National Knowledge Network (NKN) project is aimed at

XSEDE Service Provider Software and Services Baseline. September 24, 2015 Version 1.2

Interoperability between Sun Grid Engine and the Windows Compute Cluster

Grid Scheduling Dictionary of Terms and Keywords

Principles and characteristics of distributed systems and environments

PARALLEL & CLUSTER COMPUTING CS 6260 PROFESSOR: ELISE DE DONCKER BY: LINA HUSSEIN

A Taxonomy and Survey of Grid Resource Planning and Reservation Systems for Grid Enabled Analysis Environment

Grid Computing Vs. Cloud Computing

Grid Computing With FreeBSD

Resource Management on Computational Grids

Grids Computing and Collaboration

Analyses on functional capabilities of BizTalk Server, Oracle BPEL Process Manger and WebSphere Process Server for applications in Grid middleware

Symmetric Multiprocessing

Client/Server and Distributed Computing

Towards an E-Governance Grid for India (E-GGI): An Architectural Framework for Citizen Services Delivery

Service Oriented Distributed Manager for Grid System

A High Performance Computing Scheduling and Resource Management Primer

CMS Tier-3 cluster at NISER. Dr. Tania Moulik

System Models for Distributed and Cloud Computing

Data Management in an International Data Grid Project. Timur Chabuk 04/09/2007

High Performance Computing. Course Notes HPC Fundamentals

Status and Integration of AP2 Monitoring and Online Steering

So#ware Tools and Techniques for HPC, Clouds, and Server- Class SoCs Ron Brightwell

Anwendungsintegration und Workflows mit UNICORE 6

HPC Wales Skills Academy Course Catalogue 2015

Deploying a distributed data storage system on the UK National Grid Service using federated SRB

G-Monitor: Gridbus web portal for monitoring and steering application execution on global grids

Distributed Systems LEEC (2005/06 2º Sem.)

Grid Sun Carlo Nardone. Technical Systems Ambassador GSO Client Solutions

HPC-related R&D in 863 Program

An Efficient Use of Virtualization in Grid/Cloud Environments. Supervised by: Elisa Heymann Miquel A. Senar

KNOWLEDGE GRID An Architecture for Distributed Knowledge Discovery

GT 6.0 GRAM5 Key Concepts

NorduGrid ARC Tutorial

An Introduction to Virtualization and Cloud Technologies to Support Grid Computing

IBM Platform Computing : infrastructure management for HPC solutions on OpenPOWER Jing Li, Software Development Manager IBM

A Service for Data-Intensive Computations on Virtual Clusters

GridWay: Open Source Meta-scheduling Technology for Grid Computing

Classic Grid Architecture

How To Visualize Performance Data In A Computer Program

MEng, BSc Applied Computer Science

Collaborative & Integrated Network & Systems Management: Management Using Grid Technologies

SLA BASED SERVICE BROKERING IN INTERCLOUD ENVIRONMENTS

Agenda. HPC Software Stack. HPC Post-Processing Visualization. Case Study National Scientific Center. European HPC Benchmark Center Montpellier PSSC

How To Understand The Concept Of A Distributed System

Distribution transparency. Degree of transparency. Openness of distributed systems

HPC Software Requirements to Support an HPC Cluster Supercomputer

Globus Striped GridFTP Framework and Server. Raj Kettimuthu, ANL and U. Chicago

GridSolve: : A Seamless Bridge Between the Standard Programming Interfaces and Remote Resources

Developing a Computer Based Grid infrastructure

Building Platform as a Service for Scientific Applications

Introduction. MCSN N. Tonellotto Complements of Distributed Enabling Platforms

Part I Courses Syllabus

HADOOP, a newly emerged Java-based software framework, Hadoop Distributed File System for the Grid

Virtual machine interface. Operating system. Physical machine interface

16th International Conference on Control Systems and Computer Science (CSCS16 07)

Managing Complexity in Distributed Data Life Cycles Enhancing Scientific Discovery

CNR-INFM DEMOCRITOS and SISSA elab Trieste

Scientific and Technical Applications as a Service in the Cloud

The ENEA-EGEE site: Access to non-standard platforms

LSKA 2010 Survey Report Job Scheduler

Parallel Programming Survey

PRIMERGY server-based High Performance Computing solutions

Overview of HPC Resources at Vanderbilt

The GENIUS Grid Portal

Basic Scheduling in Grid environment &Grid Scheduling Ontology

Multicore Parallel Computing with OpenMP

Transcription:

HPC and Grid Concepts Divya MG (divyam@cdac.in) CDAC Knowledge Park, Bangalore 16 th Feb 2012 GBC@PRL Ahmedabad 1

Presentation Overview What is HPC Need for HPC HPC Tools Grid Concepts GARUDA Overview Grid Middleware 16 th Feb 2012 GBC@PRL Ahmedabad 2

What is HPC "High-Performance Computing," or HPC, is the application of "supercomputers" to computational problems that are either too large for standard computers or would take too long. A HPC system, is essentially a network of nodes, each of which contains one or more processing chips, as well as its own memory. These nodes are Symmetric Multi- Processors (SMP) units High-performance computing (HPC) uses parallel processing for running advanced application programs efficiently, reliably and quickly. 16 th Feb 2012 GBC@PRL Ahmedabad 3

HPC Architecture Cluster services Portal server Job Scheduler (LRM) SMP box/node having many CPUs 16th Feb 2012 GBC@PRL Ahmedabad 4

Performance of HPC Clusters Performance of HPC is measured in terms of FLOPS ( Floating-point Operations per Second). 1FLOP = Number of Instructions per Cycle X Clock Speed in Hertz X Number of CPUs HPC term applies especially to systems that function above a teraflop or 10 12 FLOPS 16 th Feb 2012 GBC@PRL Ahmedabad 5

Paradigms of Parallel Programming on HPC Types of Applications PThreads, OpenMP on SMP/Nodes MPI/PVM on Cluster i.e across nodes 16 th Feb 2012 GBC@PRL Ahmedabad 6

Need for HPC Applications Bio Chemistry: Protein Folding Chemistry: Materials Research Physics: Simulation of Heavenly bodies Environment Modeling: Earthquake, Weather Industry: Product Prototyping, Models to Calculate Risk, Processing seismographic data 16 th Feb 2012 GBC@PRL Ahmedabad 7

C-DAC s HPCC Software Products C-MPI : Optimized implementation of MPI for Cluster of Multi Processors (CLUMPS). Both point-to-point and collective calls have been optimized. Effectively uses both shared and distributed memory of CLUMPS. C-PFS : Parallel File System Provides MPI-IO file system interface to parallel applications F90IDE : Integrated Development Environment for Fortran 77/90 that includes compiler, debugger, profiler, source code browser and Fortran 77 to F90 convertor. PCF90: An automatic parallelizing compiler for Fortran for SMP based architecture. DIViA: Parallel program correctness and performance debugger. Detects communication bottlenecks and supports message debugging. PARMON: Cluster monitoring tool. Monitors the cluster as a unified resource. Provides Web Interface for monitoring over internet. RMS: Resource Management Software for effective load balancing and Load Scheduling on clusters. 16 th Feb 2012 GBC@PRL Ahmedabad 8

Grid Concepts 16 th Feb 2012 GBC@PRL Ahmedabad 9

What is Grid computing Definition of Grid by Ian Foster and Carl Kesselman A computational grid is a hardware and software infrastructure that provides dependable, consistent, pervasive, and inexpensive access to high-end computational capabilities. Three-point Check list given by Ian Foster,, Father of Grid Computing coordinates resources that are not subject to centralized control using standard, open, general-purpose protocols and interfaces to deliver nontrivial qualities of service. Reference: Book on The Grid: Blueprint for a New Computing Infrastructure. Grid puts one abstraction layer above High Performance Computing System(s), facilitating co-ordinated and distributed sharing of resources across multiple geographical and administrative domains. 16 th Feb 2012 GBC@PRL Ahmedabad 10

Origin of Grid computing Vision: The central idea is that computing should be as reliable, pervasive and transparent as a utility Information or computation power should be delivered on demand. (Apart from type and location) Origin: Conceived by academic and research communities. Internet computing grew from communication needs and GC originated from the needs of the scientific community. Create a dynamic computing environment for sharing resources and results Scale to accommodate petabytes of data, and teraflops of computing power, and keep costs down 16 th Feb 2012 GBC@PRL Ahmedabad 11

Popular Grids 16th Feb 2012 GBC@PRL Ahmedabad 12

What is Grid architecture? Grid Architecture can be described as the layers of building blocks, where each layer has a specific function, to accomplish Grid Computing Infrastructure. 16 th Feb 2012 GBC@PRL Ahmedabad 13

Components of Grid Nagios, Monalisa Ganglia,GridICE Bioinformatics,Disaster management Applications Grid Portals, PSE s, API s Access Methods Debuggers, Compiler IDE, Profilers, Workflow tools SRB/SRM, Visualization, softwares Globus/ Glite/ Unicore, Torque/Loadleveler/SGE, Gridway/MOAB/Condor-G, GSI, Grid services, MPICH & MPICH-G2 Program Development Environment Storage & Visualization Scheduler, Middleware & Security Monitoring & Management CPU, Satellite terminals, Telescopes Computational Resources & devices GARUDA, NKN and Internet links Network/Communication Fabric 16 th Feb 2012 GBC@PRL Ahmedabad 14

GARUDA Overview 16 th Feb 2012 GBC@PRL Ahmedabad 15

Objectives Motivation: To Collaborate on Research and Engineering of Technologies, Architectures, Standards and Applications in Grid Computing. Garuda Grid is targeted at providing a facility for the scientific community, which would enable them to seamlessly access the distributed resources. 16 th Feb 2012 GBC@PRL Ahmedabad 16

Connectivity Summary 16 th Feb 2012 GBC@PRL Ahmedabad 17

GARUDA Backbone A ccess Terminal Submit No de/ Grid Head node Generic User Internet Access Access Terminal LAN Port Submit No de/ Grid Head node Access Termi nal Swit ch Local User C-DAC, Bangalore Swit ch Head Node Port Head No de Hea d Node Head No de Database LAN M P L S Access Local User Comp ut e Nodes Head No de Head Node Database Head No de Local User C-DAC, Pune LAN Port A ccess Terminal Head No de Port Swit ch Hea d Node Switch Partner without resources Teles cope Head No de Grid Hea d node Database Storage A ccess Terminals 16th Feb 2012 Partner with resources GBC@PRL Ahmedabad 18

National Knowledge Network NKN 16 th Feb 2012 GBC@PRL Ahmedabad 19

GARUDA Grid: Architecture Grid-Enabled Applications Resource Enabler & Monitoring CLI Federated Information Server Access Portal Workflow tool Grid PSE Visualization Job Scheduler WSRF+GT4 + other Services +Cloud S/W Virtualization support Grid Programming & Development Environment Grid Security and High-Performance Grid Networking Data Grid NKN CDAC Resource centers Non-Research Organizations Research Organizations Computing Resources and Virtual Organizations Educational institutions Computing Centers Resources Security Middleware Resource Management User Environments Programming Environments Data Grid Grid Applications 16 th Feb 2012 GBC@PRL Ahmedabad 20

GARUDA S/W Architecture Management, Monitoring & Accounting Paryaveekshanam GARUDA Information Service Ganglia Nagios GARUDA Accounting GARUDA Resources Compute, Data, Storage, Scientific Instruments, Application Specific Software,.. Resource Mgmt & Scheduling GridWay Meta-scheduler Resource Reservation Torque, Load Leveler Globus 4.x (WS Components) Access Methods Access Portal Problem Solving Environments Cmd line interface Visuvalization gateways workflows Security Framework IGCA Certificates MyProxy VOMS 16 th Feb 2012 GBC@PRL Ahmedabad 21

GARUDA : Garuda Partner Resources Institutions Space Application Centre Indian Institute of Science Raman Research Institute Institute of Mathematical Sciences Madras Institute of Technology Indian Institute of Technology Jawaharlal Nehru University Location Ahmedabad Bangalore Bangalore Chennai Chennai Delhi Delhi Institute of Genomics and Integrative Biology Indian Institute of Technology University of Hyderabad Indian Institute of Technology Physical Research Laboratory Institute of Microbial Technology University of Pune Delhi Guwahati Hyderabad Kharagapur Ahmedabad Chandigarh Pune 16 th Feb 2012 GBC@PRL Ahmedabad 22

Grid Middleware 16 th Feb 2012 GBC@PRL Ahmedabad 23

What is a Grid Middleware Grid Middleware is a layer of S/W to enforce all the properties of: Scalability, Transparency, Heterogeneity, Fault Tolerance and Security of the Grid below the application layers. It provides an uniform interface of the Grid to users and handles all the complexity generated due to heterogeneous systems. Middleware S/W is a layer between grid applications and low level functionality of grid. 16 th Feb 2012 GBC@PRL Ahmedabad 24

Popular Grid Middlewares GT - Globus Toolkit (Argonne National Laboratory, Chicago) GRIA Grid Resources for Industrial Application (University of Southampton ) Moab Grid Suite Cluster Resources Inc. Condor Project grid computing engine by Univ. of Wisconsin Gridbus Univ. of Melbourne, for e-business and e-science Unicore Uniform Interface to Computing resources - Univ. of Virginia, for NPACI (National Partnership for Advanced Computational Infrastructure.edu) NorduGrid Middleware ARC (Advanced Resource Connector) Legion Univ. of Virginia Glite - LightWeight Middleware for Grid computing( EGEE) 16 th Feb 2012 GBC@PRL Ahmedabad 25

Core functionalities of Grid Middleware 16 th Feb 2012 GBC@PRL Ahmedabad 26

Middleware functionalities Security Job Management Data Management Information Management 16 th Feb 2012 GBC@PRL Ahmedabad 27

Grid Middleware: Security User/ Resource Authentication User/ Resource Authorization Access Control Lists User and Resource Policies Virtual Organization 16 th Feb 2012 GBC@PRL Ahmedabad 28

Grid Middleware: Job Management Submission, Status Query, Cancel & Destroy, Getting Output & Error Support an open Job Description Language RSL, JDL, JSDL Transferring input/output data from/to remote source/destination Support Serial/ Parallel Jobs (Heterogeneous & Homogeneous) Integration with all Local Resource Managers 16 th Feb 2012 GBC@PRL Ahmedabad 29

Grid Middleware: Data Management Two Basic Categories of Data Management Data Movement Secure Robust Efficient Third party movement Data Replication One or more copies or replicas Survive loss Easy availability Reduce access latency, increase robustness, scalability and performance for distributed applications. 16 th Feb 2012 GBC@PRL Ahmedabad 30

Grid Middleware: Information Management System information is critical to operation of the grid and construction of applications How does an application determine what resources are available? What is the state of the computational grid? How can we optimize an application based on configuration of the underlying system? We need a general information infrastructure to answer these questions 16 th Feb 2012 GBC@PRL Ahmedabad 31

Examples of Useful Information Characteristics of a compute resource IP address, software available, hostname, nodes available, OS version, load, softwares, libraries & their licenses. Characteristics of a network Bandwidth and latency, protocols Characteristics of the Middleware infrastructure Hosts, local resource managers 16 th Feb 2012 GBC@PRL Ahmedabad 32

Thank You 16 th Feb 2012 GBC@PRL Ahmedabad 33