Section: Partnerships and Cooperations

International Initiatives

• Title: Thoth – An Automatic Dynamic Binary Parallelisation System

• International Partner (Institution - Laboratory - Researcher):

  • Egypt-Japan University of Science and Technology - Prof. Ahmed ElMahdy.

• Dates: 2016–2017

With the current global trend towards utilizing cloud computing and smart devices, executing the same application across becomes a necessity. Moreover, parallelism is now abundant with various forms that include thread- and data-parallel execution models. Such diversity in ISA and explicit parallelism makes software development cost prohibitive, especially for natively optimized binaries. This project leverages dynamic binary translation technology to provide for exploiting the underlying parallel resources without the need of having the source code of the application. In particular the project integrates low overhead dynamic profiling, novel OSR parallel de-optimization and a retargetable parallelization modules to allow for dynamic parallelization of binaries.

• Title: Profiling and specialization for locality

• International Partner (Institution - Laboratory - Researcher):

  • Universidade Federal de Minas Gerais (Brazil) - DCC - Fernando Magno Quintão Pereira

• Start year: 2015

• See also: https://team.inria.fr/pacap/prospiel/

• The PROSPIEL project aims at optimizing parallel applications for high performance on new throughput-oriented architectures: GPUs and many-core processors. Traditionally, code optimization is driven by a program analysis performed either statically at compile-time, or dynamically at run-time. Static program analysis is fully reliable but often over-conservative. Dynamic analysis provides more accurate data, but faces strong execution time constraints and does not provide any guarantee. By combining profiling-guided specialization of parallel programs with runtime checks for correctness, PROSPIEL seeks to capture the advantages of both static analysis and dynamic analysis. The project relies on the polytope model, a mathematical representation for parallel loops, as a theoretical foundation. It focuses on analyzing and optimizing performance aspects that become increasingly critical on modern parallel computer architectures: locality and regularity.