RSEEE & WHPC Meeting - East Anglia 2026
22 April 2026 · JIC Conference Centre · Norwich
These are the submitted talk abstracts for the event programme.
Speaker: Tom Meltzer (University of Cambridge)
The free and open-source MPI debugging landscape is limited, and paid tools are not consistently available across HPC systems. mdb is presented as a free, portable MPI debugger built around familiar gdb-style workflows, with support for C, C++, Fortran, and Rust and compatibility with multiple debugger backends. The talk explains how mdb can help researchers reuse existing debugging knowledge across different environments.
Speaker: Martin Vickers (John Innes Centre)
At the John Innes Centre, informatics support grew organically around immediate needs, enabling rapid progress but not long-term software sustainability. Over the past four years, the team has introduced RSE roles, clarified ownership, and reduced technical debt to shift from reactive maintenance to proactive development. This talk reflects on that transition and outlines practical steps to ensure that the apparent zero-cost speed of AI-enabled development does not become tomorrow's maintenance burden.
Speaker: Sathish Thiyagarajan (Earlham Institute)
This session presents nanalogue-gui, a sequencing data processing application with AI chat features that lets users interact with genomics data in plain English through a point-and-click interface. It supports hosted and local model providers, including on-premises workflows for sensitive data, and executes AI-generated code in a secure sandbox to protect files and maintain reproducibility.
Speaker: Navirah Kamal (University of Cambridge)
Research computing systems consume significant resources, but quantifying energy use and carbon footprint at user and job level remains difficult. The Green Algorithms Dashboard addresses this by using scheduler data to provide fine-grained monitoring across users, groups, departments, and institutions. This talk presents the platform's architecture, privacy-first local deployment model, and insights from early pilots, alongside planned support for GPU-heavy workloads and broader scheduler integration.
Speaker: Will Redding (JIC)
Provisional lightning talk exploring methods and findings related to merging fast neutron induced deletions in pea genomics research.
Speaker: Chris Edsall (University of Cambridge)
Cambridge is host to a new HPC resource, a follow-on system to Dawn and one of the four phase 1 National Compute Resources. We describe Zenith, a 72 node, 576 GPU system built around AMD MI355x GPUs. As well as the system details, this talk covers access routes, recommended programming models, and the planned software stack.
Speaker: Lisa Crossman (UEA/QIB)
As genomics datasets continue to expand, execution speed and memory management in interpreted languages such as Python and R are increasingly becoming computational bottlenecks. This talk presents microBioRust, an open-source library written in Rust and optimized for microbial genomics, together with Python interoperability through microbiorust-py. Benchmarking has shown major speed improvements compared with Biopython for key tasks, and the session invites broader community collaboration for robust testing and ecosystem growth.