D9.6: Final plan to disseminate and use the foreground knowledge

In this deliverable, we examined the overall heterogeneous computing market, trends, and players in order to determine potential effective approaches and strategies that form good bases for exploitation and dissemination of results achieved during REPARA work. The heterogeneous processing and Continue reading D9.6: Final plan to disseminate and use the foreground knowledge

D9.3: Report on Initial Project Web Site

This report describes the initial website design, including the public website (available at http://www.repara-project.eu) and the project internal website, as well as other internet based dissemination means. Results reported in this deliverable will influence the approach for dissemination and public Continue reading D9.3: Report on Initial Project Web Site

D6.4: Performance and energy monitoring library

Work Package 6 is focusing on run-time engines that manage software running on parallel heterogeneous computing systems, which includes techniques for measuring performance and energy consumption of the managed software and the integration of these techniques into the run-time engine. Therefore, we designed Continue reading D6.4: Performance and energy monitoring library

D7.1: High-level quantitative performance and energy models

In this report we describe a method for deriving high-level quantitative prediction models for making low-fidelity predictions on performance, power, and energy consumption when a software component is mapped to a specific processing element. These predictions are low-fidelity as the transformation specific Continue reading D7.1: High-level quantitative performance and energy models

D4.3: Source Code Transformations for Coarse Grained Parallelism

The REPARA project aims to help the transformation and deployment of new and legacy applications in parallel heterogeneous computing architectures while maintaining a balance between application performance, energy efficiency and source code maintainability. This report describes, source code transformations for coarse grained parallelism. The Continue reading D4.3: Source Code Transformations for Coarse Grained Parallelism

D2.4: Algorithm Constraining and Selection Techniques

This report describes algorithm constraining and selection techniques. The presented techniques allow generating multi-versioned kernels implementations for a selected kernel in a source code in order to choose the most promising one in runtime. The report introduces the workow and the algorithms used. The Continue reading D2.4: Algorithm Constraining and Selection Techniques