The advent of manycore and heterogenous systems with accelerators requires that programmers understand how to design, write, and debug parallel programs effectively.

The increased complexity of multi-threaded and/or accelerator-based parallel programming requires more insight into program behavior and necessitates the use of tools that can support programmers in migrating existing software to multi-/manycore and accelerator platforms as well as in writing new parallel software. Programmers need increasingly sophisticated methods for instrumentation, measurement, analysis and modeling of applications.

To help software developers embrace the complex platform landscape, we need both sophisticated tools to support parallel software development, and an extensible tool infrastructure that facilitates the integration of existing and new tools, and simplifies information exchange and access.