Message Passing Programming (P01655)

? Credit Points : 10 ? SCQF Level : 11 ? Acronym : PHY-P-HPCMesPas

The world's largest supercomputers are used almost exclusively to run applications which are parallelised using Message Passing. This course covers all the basic knowledge required to write parallel programs using this programming model, and is directly applicable to almost every parallel computer architecture.

Parallel programming by definition involves co-operation between many separate tasks to solve a common problem. The programmer has to define the tasks that will be executed by each of the parallel processors, and also how these tasks are to synchronise and exchange data with one another. In the message-passing model, each task is a separate operating system process; these processes communicate and synchronise by explicitly sending each other messages. All these parallel operations are performed via explicit calls to some message-passing library that is entirely responsible for interfacing with the physical communication network linking the processors together.

This course uses the de facto standard for message-passing libraries: the Message Passing Interface (MPI). The syllabus covers: the message-passing model; compiling MPI programs; point-to-point communication; non-blocking operations; derived datatypes; virtual topologies; collective communication; high-level MPI design issues. Each topic is illustrated with a small practical programming example. At the end of the course there is a more substantial Case Study exercise in parallel image processing which uses all the techniques covered in the lectures.

Entry Requirements

? Pre-requisites : Only postgraduates or final year MPhys students with suitable computational background, subject to space restrictions and agreement with relevant Programme Coordinator.

Subject Areas

Home subject area

Postgraduate (School of Physics), (School of Physics, Schedule Q)

Delivery Information

? Normal year taken : Postgraduate

? Delivery Period : Block 5 and beyond

? Contact Teaching Time : 20 hour(s) per week for 1 weeks

All of the following classes

Type	Day	Start	End	Area
Lecture	Tuesday	09:00	17:00	KB

Summary of Intended Learning Outcomes

On completion of this course students should be able to:

- Understand the message-passing model in detail.
- Implement standard message-passing algorithms in MPI.
- Debug simple MPI codes.
- Measure and comment on the performance of MPI codes.
- Design and implement efficient parallel programs to solve regular-grid problems.