Skip to main content Skip to navigation

PX457 High Performance Computing in Physics

Lecturer: Nick Hine
Weighting: 15 CATS

The module will address the increased use of computer simulation and data analysis on high performance computers in all fields of computational physics and other sciences. Computing skills are greatly valued across science and beyond and we encourage students to go as far as they can to develop such skills.

Aims:
To explain the methods used in computer simulations and data analysis on high performance computers, for research in all fields of computational physics and other sciences.

Objectives:
By the end of the module, students should be able to:

  • Identify and correct common inefficiencies in both serial scientific computer codes
  • Write a parallel program using shared-memory or message passing constructs in a physics context, and to write a simple GPU accelerated program
  • Choose an appropriate programming paradigm and identify sources of performance bottlenecks and parallelisation errors in parallel computer programs and understand how these relate to the computer architecture
  • Process very large datasets with appropriate tools, and use machine-learning to efficiently extract simple functional forms describing sparse data

Syllabus:

Programming for efficiency. What does a compiler do that an interpreter cannot? The mechanics of linking C and Python. Modern cache architectures and CPU pipelining. Avoiding expensive and repeated operations. Compiler optimisation flags.

Introduction to parallel computing. Modern HPC hardware and parallelisation strategies. Analysing algorithms and codes to identify opportunities for parallelism.

Shared memory programming. The OpenMP standard. Parallelisation using compiler directives. Threading and variable types. Loop and sections constructs. Program correctness and reproducibility. Scheduling and false sharing as factors influencing performance.

Distributed memory programming. The MPI standard for message passing. Point-to-point and collective communication. Synchronous vs asynchronous communication. MPI communicators and topologies.
GPU programming. CUDA vs OpenCL. Kernels and host-device communication. Shared and constant memory, synchronicity and performance. GPU coding restrictions.

Limitations to parallel performance. Strong vs weak scaling. Amdahl’s law. Network contention in modern many-core architectures. Mixed mode OpenMP+MPI programming.

“Big Data” in physics: handling very large datasets. Examples derived from astronomy and particle physics.

Commitment: 25 Lectures + 5 Laboratory Sessions

Assessment: Assignments (100%)

Recommended Texts:
R Chandra et. al,. Parallel Programming in OpenMP , Morgan Kaufmann, P Pacheco, Parallel Programming with MPI Morgan Kaufmann
M Quinn, Parallel Programming in C with MPI and OpenMP McGraw-Hill
D Kirk and W Hwu, Programming Massively Parallel Processors Elsevier