Special issue: Euro‐Par 2016

This special issue of Concurrency and Computation: Practice and Experience contains revised and extended versions of selected papers presented at the conference Euro-Par 2016. Euro-Par—the European Conference on Parallel Computing—is an annual series of international conferences dedicated to the promotion and advancement of all aspects of parallel and distributed computing. Euro-Par covers a wide spectrum of topics from algorithms and theory to software technology and hardware-related issues, with application areas ranging from scientific to mobile and cloud computing. The major part of the Euro-Par audience consists of researchers in academic institutions, government laboratories, and industrial organisations. Euro-Par 2016, the 22nd conference in the Euro-Par series, was held in Grenoble, France. It was organised by Inria, Université Grenoble-Alpes, and IUT 2 Grenoble. Twelve broad topics were defined and advertised, covering a large variety of aspects of parallel and distributed computing. The call for papers attracted a total of 176 submissions. The submitted papers were reviewed at least 3 and, in most cases, 4 or even more times (4 reviews on average). A total of 47 papers were finally accepted for publication. This makes a global acceptance rate of 26.7 %. The authors of accepted papers came from 20 countries, with the 4 main contributing countries—France, the United States, Germany, and Spain—accounting for a bit more than half of them. Based on the results of the reviews and a majority opinion of the respective topic programme committees, a number of papers were recommended for this special issue. The authors who gave a convincing talk were contacted at the conference and invited to submit revised and extended versions of their papers. These new versions were given to 3 reviewers; 2 had previously reviewed the conference version, the third had not. Eventually, five papers were accepted for publication. This year, four Euro-Par topics are represented. Topic 3 on Scheduling and Load Balancing is represented by the paper Controlling the correlation of cost matrices to assess scheduling algorithm performance on heterogeneous platforms.1 The authors Louis-Claude Canon, Pierre-Cyrille Héam, and Laurent Philippe consider the problem of allocating independent tasks to a collection of different machines in an optimizing manner. Costs are assessed via the generation of matrices that advise on the cost of a specific task on a specific machine. The authors propose a new assessment metric comparing the uniform cost matrix and a cost matrix using task and machine correlations. Two generation methods are proposed, one of them new, the other a modification of an existing one, and their effect on performance evaluation heuristics from the literature is studied. One reviewer pointed out that, despite the numerous required mathematical proofs, the authors do an admirable job of making each one accessible. Topic 4 on High-Performance Architectures and Compilers is represented by 2 papers. In their paper Piecewise holistic autotuning of parallel programs with CERE,2 the authors Mihail Popov, Chadi Akel, Yohan Chatelain, William Jalby, and Pablo de Oliveira Castro describe how to autotune codelets that have been produced by the Codelet Extractor and Replayer CERE. Autotuning a set of codelets incurs lower costs than autotuning the entire program, and CERE helps in setting optimization options appropriately. The good design of the tool and the informative role of the experiments were appreciated in reviewing. The authors Juan Manuel Martinez Caamaño, Manuel Selva, Philippe Clauss, Artyom Baloian, and Willy Wolff follow, with their tool APOLLO, a similar approach in the polyhedron model for loop optimization. In their paper Full runtime polyhedral optimizing loop transformations with the generation, instantiation and scheduling of code-bones,3 they propose a compile-time generation of so-called code-bones that are then, at run time, instantiated and assembled to target code. In the optimization of the target code, the popular polyhedral tools PluTo and CLooG are being employed. Since the code-bones have an increased flexibility, the results show improvements over APOLLO's predecessor VMAD. Topic 5 on Parallel and Distributed Data Management and Analytics is represented by the paper A flexible I/O arbitration framework for netCDF-based Big Data processing workflows on high-end supercomputers.4 The authors Jianwei Liao, Balazs Gerofi, Guo-Yuan Lien, Takemasa Miyoshi, Seiya Nishizawa, Hirofumi Tomita, Wei-keng Liao, Alok Choudary, and Yutaka Ishikawa study a direct communication framework designed for complex workflows that eliminates unnecessary file I/O between job components. Their idea is to inject an I/O arbitration layer that provides direct parallel data transfer. Users are also given the opportunity to specify the desired data transfer pattern in a file. One reviewer noted that the engineering presented is of great interest to the community. Finally, Topic 9 on Multicore and Manycore Parallelism is represented by the paper A framework for dense triangular matrix kernels on various manycore architectures,5 authored by Ali Charara, David Keyes, and Hatem Ltaief. They present a new high-performance framework for architecture-oblivious dense triangular BLAS kernels—triangular matrix-matrix multiplication (TRMM) and triangular solve (TRSM)—on various manycore architectures. Compared to the conference version, this framework is enhanced further with the availability of customized CUDA kernels and a multiple-GPU implementation with almost linear scalability. The reviewers appreciated the high quality and scientific soundness of the treatise. Concluding this preface, we would like to thank Prof Geoffrey Fox, editor-in-chief of Concurrency and Computation: Practice and Experience, for his support of this special issue. We would also like to thank our peers who assisted us in reviewing the papers and helped strengthen the final