Modern HPC architectures give us access to amazingly powerful machines, yet many of us still run jobs on our desktops.
Although not every problem is suitable for parallelisation, many of them are. If we could learn more about what's available and how people have used it, we could get much of our work done a lot faster and more efficiently and perhaps even open new ways of approaching design and operation problems.
The first day is focussed on users of existing codes, with demonstrations and tutorials. The second is aimed at the writers of programs, looking at how codes can be speeded up with case studies from accelerator and other applications. The third is aimed to involve project/group leaders and other interested participants, and will discuss possibilities for new initiatives, aiming to form some collaborations that will lead to proposals. These three groups are, of course, not distinct - most of us wear at least two hats - and people are encouraged to attend for the day or days they think they will find most useful.
Full details are given (the programme is still evolving) on the Registration page (link opens in a new window) (there is no fee).
