Debugging at Scale with Allinea DDT

Abstract

As core counts in HPC clusters grow, almost every application user is trying to run software at higher scale than before. It is not always an easy task and can end in failure as the limitations of existing software are discovered. Users and developers quickly find new (and old) bugs as scale increases: software can be complex when it seeks to use more threads or processes to exploit the hardware. In this document, we show how using a debugger at the scale of the bug is the most effective way to tackle parallel software problems today. We introduce Allinea DDT – the world’s only scalable parallel debugger – and show how it is fast, capable, and lets you debug your parallel or multithreaded application, no matter how big or small a system you use, easily. Allinea DDT has been setting standards for usability for many years and has torn up scalability records. It is used on the world’s largest systems – debugging over 220,000 processes simultaneously in some cases. Bugs can be fixed easily for all developers – not just those with extreme scale – by using Allinea DDT at your scale.