SCIENCE

Addressing large scale debugging with MPC and CUDA

Allinea Software http://www.allinea.com has signed a further collaboration agreement with CEA, the French government-funded technological research organisation http://www.cea.fr/english_portal.   

The aim of the collaboration to date has been to develop enhancements to Allinea Software’s market-leading Allinea DDT (Distributed Debugging Tool) for next generation hybrid and “many-core” computer systems.  Based on the work carried out in phase one, the two organisations are seeking to address a major challenge and to offer a solution for large scale systems.

Dr. David Lecomber, CTO at Allinea Software, explains: “Research organisations require ever more powerful computational resources, and so their software needs to run on ever larger numbers of processors. This increases the complexity and risks of errors in software development; Allinea is one of the only suppliers of tools that can handle this complexity and identify errors before they hold up important research.

Based on all of the knowledge we acquired from the first project with CEA, this new collaboration intends to address the concerns and support the aims of all HPC developers worldwide. A key aim is clearly to provide a solution now and in the future for MPC unified parallel runtime (see:http://mpc.sourceforge.net/) large scale debugging.”

The focus of phase two is to make debugging tools both portable and easy to use for large scaledebugging well over 100,000 cores. Allinea DDT's interface is intuitive at every scale of parallelism, and itsarchitecture has already been shown to scale well to existing large systems.  Another important aim is to improve the previous work carried out on CUDA debugging by adding more features to address the CUDA architecture and fully exploit the enhancements which NVidia now offers.

Pierre Leca, head of Simulation and Information Sciences Department, CEA, comments: “We’re verypleased with the results of our collaboration with Allinea Software over the past year and we want this to continue.  They are clearly an active, technology leader, with the right product already available, and unrivalled expertise in parallel performance. 

Working together, we can address the growing numbers of cores in clusters and within individual nodes to take advantage of optimizations that are feasible for these systems. This will ensure that developers can use a tool that scales well on large many-core clusters both in terms of performance and in ease of use.”

This next phase of collaboration with CEA is a further illustration of Allinea Software’s leadership in High Performance Computing (HPC) software with the most demanding  HPC sites in both Europe and the USA. 

Allinea Software will be exhibiting at the Supercomputing Conference (SC’10) in New Orleans, USA from November 15 -18, 2010  - booth #2305.