Course description: Application performance is one of the key requirements for HPC applications. However, this is one of the more difficult requirements to satisfy. Issues affecting performance often vary between different hardware and software environments. This requires performance issues to be frequently re-visited as the hardware and software environment changes.
This course will concentrate on single processor optimisation techniques.
Topics include: overview of performance programming; methodology of the optimisation cycle; designing for performance; encapsulation as an aid to performance tuning; tools for performance programming; profilers and code instrumentation; compilers and compiler optimisation; memory heirarchies; memory structures and associated optimisations.
Learning outcome: Students will gain a good practical understanding of the general issues and methodologies associated with designing building and refactoring codes to meet performance requirements.
Pre-requisites: C, C++ or Fortran.
Course overview:
Session I: Performance Optimisation
This session introduces the subject and looks at the work-flow behind performance optimisation. It also provides an overview of current microprocessor architectures concentrating on performance aspects.
Session II: Compiler based optimisation
This session looks at compiler based optimisations. What optimisations a compliler is able to perform. What might prevent a compiler from performing these optimisations and how to help the compiler do its job.
Session III: Memory optimisation
This session looks at the impact that memory utilisation and data structures can have on program performance and how to re-design data structures for increased performance.
Session IV: Object Oriented optimisation
This session looks at how Object Oriented techniques impact code performance and how they can be used to aid the optimisation process.
For a more detailed timetable, please click on the "Timetable" link from the menu on the left.
This course will concentrate on single processor optimisation techniques.
Topics include: overview of performance programming; methodology of the optimisation cycle; designing for performance; encapsulation as an aid to performance tuning; tools for performance programming; profilers and code instrumentation; compilers and compiler optimisation; memory heirarchies; memory structures and associated optimisations.
Learning outcome: Students will gain a good practical understanding of the general issues and methodologies associated with designing building and refactoring codes to meet performance requirements.
Pre-requisites: C, C++ or Fortran.
Course overview:
Session I: Performance Optimisation
This session introduces the subject and looks at the work-flow behind performance optimisation. It also provides an overview of current microprocessor architectures concentrating on performance aspects.
Session II: Compiler based optimisation
This session looks at compiler based optimisations. What optimisations a compliler is able to perform. What might prevent a compiler from performing these optimisations and how to help the compiler do its job.
Session III: Memory optimisation
This session looks at the impact that memory utilisation and data structures can have on program performance and how to re-design data structures for increased performance.
Session IV: Object Oriented optimisation
This session looks at how Object Oriented techniques impact code performance and how they can be used to aid the optimisation process.
For a more detailed timetable, please click on the "Timetable" link from the menu on the left.