Description
This group carries out activities related to performance analysis, simulation, emulation, and development of new libraries for emerging architectures. The group develops tools and models to monitor the use of resources and measure the performance in parallel and distributed HPC applications. These performance analysis tools provide detailed analysis capabilities to make it possible to understand both the application and the underlying system s behaviour, and identifying performance critical issues. In addition, the group uses the insights provided by its performance analysis methodology to propose novel methods and libraries able to efficiently run irregular workloads like sparse linear algebra kernels.
