Towards Transparent Resilience for the Chapel Parallel Language Ms Konstantina Panagiotopoulou, Heriot-Watt University The rapidly increasing number of components in modern High Performance Computing (HPC) systems provides a challenge on their resilience; predictions of time between failures on ExaScale systems range from hours to minutes. Yet, the prevalent HPC programming model today does not tolerate faults. This talk features the design and initial implementation of transparent resilience for Chapel, a parallel HPC language following the Partitioned Global Address Space (PGAS) programming model. We address cases of hardware failure on one or multiple nodes during program execution in a distributed setup, using detection and recovery mechanisms. We focus on the runtime system, particularly on the communication (GASNet) and tasking layers to address task parallelism and extend the work on library level to handle data parallelism. Ongoing work addresses integration of distributed task adoption strategies with Chapel's default data distributions. This talk summarises results and experiences from a 2-month internship with the Chapel developer's group at Cray. Biography: I'm a third year PhD student at Heriot-Watt University and member of the Dependable Systems Group. My research interests are in the design and implementation of Partitioned Global Address Space (PGAS) languages, in the context of High Performance Computing, with focus on resilience. PGAS languages implement the concept of a shared global address space, and use language constructs to distribute data structures over machines, providing the programmer with opportunities to tune data locality and enhance performance. Resilience, is one of the main challenging topics at ExaScale, and an area of gaining popularity among researchers.