Multicore and manycore architectures are nowadays pervasive in all computing fields. Yet, their intrinsic complexity creates several programming challenges when performance, portability, and power efficiency are the most important metrics to optimize. Programming complexity is further exacerbated when considering heterogeneous platforms comprising thousands of heterogeneous cores. To use such systems efficiently, algorithms must scale to a large degree of parallelism, utilize optimized data structures and synchronization mechanisms, and leverage fine-tuned parallel runtimes and frameworks capable of reducing parallel overheads.

This topic seeks to explore the programming of homogeneous and heterogeneous multicore and manycore systems. It focuses on novel research and solutions in the form of programming models, algorithms, concurrent data structures, libraries, runtime systems, and tools capable of increasing programmability and performance of multi- and manycore systems in the context of general-purpose, high-performance, and embedded parallel computing.

Focus

• Parallel programming models, and languages
• Tools and frameworks for parallel programming
• Efficient concurrent algorithms and concurrent data-structures
• Libraries and runtimes for parallelism
• Advances in performance-portability and performance-power tradeoffs
• Innovative applications and case studies
• Advances in parallel software development and software productivity
• Tools and techniques for discovering and understanding parallelism
• Performance modeling and analysis of multicore and manycore systems
• Power efficient parallel programming techniques