Continous integration for High Performance Computers Continous integration for High Performance Computers Continous integration for High Performance Computers

Workshop on Continuous Integration on High Performance Computing Platforms

A worshop on testing software on supercomputers, organised at Imperial on the 10th November 2017 and funded by the software sustainability institute. The process of continous integration, that is automatic testing of code whenever a change is made by any user, is routine in software engineering firms. However, it is generally absent in academia. As we move to the next generation of research software, increasing large open source projects are the norm. This means that validated code must remain valid. In order to ensure this, it becomes essential to automate tests which ensure use contributions to the software don't break basic functionality. This is especially true on the nations supercomputers, when as code complexity grows, the next generation of scientists becomes users instead of developers, unable to fix problems when the code does not run as expected. This workshop aims to bring together research software engineers (RSE) and academics to discuss the best techniques to implement this approach.

Material from the day, includes:

• Summary document from the day.
• Notes for steering committee and results from group execises.
• List of attendees.
• Day Plan.

Talks

• Edward Smith.
• Chris Cantwell.
• Catherine Jones.
• David Pérez-Suárez.
• Jeffrey Salmond and Krishna Kumar.