01_overview - Motivation Architectures Networks...

Info iconThis preview shows pages 1–8. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Motivation Architectures Networks Communication Parallel Numerical Algorithms Chapter 1 Parallel Computing Prof. Michael T. Heath Department of Computer Science University of Illinois at Urbana-Champaign CSE 512 / CS 554 Michael T. Heath Parallel Numerical Algorithms 1 / 49 Motivation Architectures Networks Communication Outline 1 Motivation 2 Architectures Taxonomy Memory Organization 3 Networks Network Topologies Graph Embedding 4 Communication Message Routing Communication Concurrency Collective Communication Michael T. Heath Parallel Numerical Algorithms 2 / 49 Motivation Architectures Networks Communication Why Parallel Computing? Parallel computing provides cost-effective (often only) means of meeting enormous resource demands of large-scale simulations leverages existing resources or relatively inexpensive commodity parts offers alternative when individual processor speeds ultimately reach limits imposed by fundamental physical laws Michael T. Heath Parallel Numerical Algorithms 3 / 49 Motivation Architectures Networks Communication Reasons for Caution Parallel computing should be approached with caution unstable commercial market relative lack of available software immature computing environment challenging complexity of parallel programming physical constraints (power, cooling, packaging, etc.) Nevertheless, insatiable appetite of computational scientists for ever greater computing capability has led them to embrace large-scale parallelism Michael T. Heath Parallel Numerical Algorithms 4 / 49 Motivation Architectures Networks Communication Taxonomy Memory Organization Flynns Taxonomy Flynns taxonomy : classification of computer systems by numbers of instruction streams and data streams: SISD : single instruction stream, single data stream conventional serial computers SIMD : single instruction stream, multiple data streams special purpose, data parallel computers MISD : multiple instruction streams, single data stream not particularly useful, except perhaps in pipelining MIMD : multiple instruction streams, multiple data streams general purpose parallel computers Michael T. Heath Parallel Numerical Algorithms 5 / 49 Motivation Architectures Networks Communication Taxonomy Memory Organization SPMD Programming Style SPMD (single program, multiple data): all processors execute same program, but each operates on different portion of problem data Easier to program than true MIMD, but more flexible than SIMD Although most parallel computers today are MIMD architecturally, they are usually programmed in SPMD style Michael T. Heath Parallel Numerical Algorithms 6 / 49 Motivation Architectures Networks Communication Taxonomy Memory Organization Architectural Issues Major architectural issues for parallel computer systems include processor coordination : synchronous or asynchronous?...
View Full Document

This note was uploaded on 11/06/2011 for the course CSE 494 taught by Professor Staff during the Summer '09 term at CUNY Brooklyn.

Page1 / 49

01_overview - Motivation Architectures Networks...

This preview shows document pages 1 - 8. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online