lect02 - Notes ! Accounts handed out and first assignment...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
CMSC 714 High Performance Computing Lecture 2 - Introduction http://www.cs.umd.edu/class/fall2011/cmsc714 Alan Sussman 2 CMSC 714 – F11 Alan Sussman & Jeffrey K. Hol ingsworth Notes ! Accounts handed out and first assignment probably late next week 3 CMSC 714 – F11 Alan Sussman & Jeffrey K. Hol ingsworth Last time ! Why parallel computing? – speed, cost ! Parallel computing basics – Processing elements, memory, network, disks – SIMD, MIMD, SPMD, dataflow – networks • bus, ring, tree, mesh (2D or 3D), hypercube – memory • latency and throughput (bandwidth) • shared vs. distributed (physically and logically) • UMA vs. NUMA 4 CMSC 714 – F11 Alan Sussman & Jeffrey K. Hol ingsworth Coordination ! Since parallelism in our view is processors working together to solve a problem ! Synchronization – protection of a single object (e.g., locks) – coordination of processors (e.g., barriers) ! Size of a unit of work by a processor – need to manage two issues • load balance - processors have equal work • coordination overhead - communication and synchronization – often called “grain” size - coarse grain vs. fine grain
Background image of page 1

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

View Full DocumentRight Arrow Icon
5 CMSC 714 – F11 Alan Sussman & Jeffrey K. Hol ingsworth Sources of Parallelism ! Statements – called “control parallel” – can perform a series of steps in parallel – basis of dataflow computers ! Loops – called “data parallel” – most common source of parallelism for most programs – each processor gets one (or more) iterations to perform 6 CMSC 714 – F11 Alan Sussman & Jeffrey K. Hol ingsworth Example of Parallelism ! Easy (embarrassingly parallel) – multiple independent jobs (i.e. .., different simulations) ! Scientific – dense linear algebra (divide up matrix) – physical system simulations (divide physical space) ! Databases – biggest success of parallel computing (divide tuples) • exploits semantics of relational algebra ! AI – search problems (divide search space) – pattern recognition and image processing (divide image) 7 CMSC 714 – F11 Alan Sussman & Jeffrey K. Hol ingsworth Metrics in Application Performance ! Speedup
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 01/12/2012 for the course CMSC 714 taught by Professor Staff during the Fall '07 term at Maryland.

Page1 / 5

lect02 - Notes ! Accounts handed out and first assignment...

This preview shows document pages 1 - 3. Sign up to view the full document.

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