Correctness and Debugging Workshop Series

Within modern, complex software landscapes, developers typically spend a significant proportion of their time on debugging. Studies estimate that more time is required to find bugs than to write actual code. Finding bugs and delivering correct(er) code becomes more difficult with modern HPC systems which offer unprecedented concurrency: The systems become heterogeneous, i.e. process code differently, bugs become non-deterministic. Different programming models are often used in combination, complicating questions of correctness and debugging further.

This workshop series gives a hand-on introduction to modern tools for debugging and correctness checking. It consists of the following events.

• 05 November, 1pm-5pm (online): Debugging Principles
• 06 November, 1pm-5pm (online): Linaro DDT
• 26 November, 1pm-5pm (online): Compiler Sanitizer
• 27 November, 1pm-5pm (online): Correctness checker
• 8 December: In-Person Wrap Up Workshop at UCL

Participants are expected to attend all sessions.

• Each participant must bring their own laptop.
• Each particpant should bring some research code to play with and analyze. This should be in a compiled language (e.g. C/C++ or Fortran). Pure Python code will not work as it does not produce compiled code.
• Participants should have access to the Durham test beds. More details on this will be communicated.

Accommodation and travel information for in-person day

Due to the HAI-End project, there are some limited travel bursaries available for non-UCL participants. Please contact tobias.weinzierl@durham.ac.uk directly for questions regarding this support. Money is available to postgraduate research students, research technology professionals and junior academics on a FCFS basis.

For accommodation we recommend the following hotels around the UCL Campus:

• Tavistock Hotel
• Grange Clarendon
• City sleeper at Royal National Hotel

There are no preferential rates available for these hotels.

Acknowledgements

In collaboration with the UCL Advanced Research Computing Centre, the EPSRC CCMI CDT, VI-HPS and the Institute for Data Science (Durham University). The organisers appreciate support through the UKRI Digital Research Infrastructure Programme [grant number UKRI/ST/B000293/1 (HAI-End)].