When: 12 - 14 September 2018, 09:30 - 17:30
Where: Hartree Centre, Sci-Tech Daresbury, Cheshire, UK
Best for: Research software engineers, postgraduate students and scientists with computing intensive workloads.
Prerequisites: Basic skills in C/C++ or Fortran programming; Some experience in using a Linux HPC environment.
Summary: This free course brings together Intel and Hartree Centre parallel software experts to teach how to write efficient OpenMP code and use Numerical Libraries with the help of Intel Compiler and the Intel performance analysis tools. It is essential to write efficient parallel code to extract the maximum performance from each processor, as the number of cores and the SIMD capabilities of the Intel processors increases. OpenMP is a directive-based open standard that allows annotating your code with work sharing, SIMD and task-based constructs to achieve optimal node performance.
The course will provide theory, hands-on, and the possibility to bring your own code to analyse the performance and optimise it with the help of Hartree Centre experts.
|Day 1||Introduction to OpenMP|
Analysing performance with Intel VTune
Analysing performance with Intel Advisor
Bring your own code* and work with Hartree/Intel experts to optimise it
Visit the Hartree Centre visualisation and computing facilities
* We encourage participants to bring the code from the domain which they are more familiar with/interested in optimise. If you do not have a code to use, we will provide Hartree Centre software to perform analysis and optimisations.