This workshop brought together 36 researchers from across the DiRAC and wider computational science community, to identify and review common challenges and techniques in contemporary research and gain pointers on how to navigate the rapidly evolving hardware landscape. 

Highlights included presentations from experts on Performance Portability and Use of Accelerators (accompanied by Case Studies led by experienced practitioners); Open Science and Reproducibility; Algorithms and Code Sustainability; Upskilling and Training; and finally on Building Communities. The talks were interspersed with lively group discussions exploring themes such as GPUs and me, What Training do I need?, and Who is my Community? 

The intended outcome was to identify and nurture groups of researchers from different projects but with interests or issues in common, ready to share their experience and foster ongoing knowledge exchange by building sustainable and inclusive communities; judging from the feedback received some useful steps in this direction were taken. 

Session Presentations

Algorithm Choice for Modern Hardware – Jamie Quinn

My confused and incomplete journey from CPUs to GPUs – Mark Hannam

Building a (CCP) community – Eugene Lim

Case Study: Black Holes, JAX and GPUs – James Nightingale

Empathetic Software Development – Ed Bennett

Reproducibility and Open Science: Why and How? – Ed Bennett

Performance and performance-portability in ExaGRyPE – Tobias Weinzierl

The DiRAC-based organisers were:
Simon Burbidge (DiRAC RSE Team Lead) 
Simon Hands (DiRAC Community Development Director) 

together with local expertise from 
Katy Clough (QMUL)
Biagio Lucini (QMUL and DiRAC Technical Directorate)