No Arabic abstract
Cloud computing refers to maximizing efficiency by sharing computational and storage resources, while data-parallel systems exploit the resources available in the cloud to perform parallel transformations over large amounts of data. In the same line, considerable emphasis has been recently given to two apparently disjoint research topics: data-parallel, and eventually consistent, distributed systems. Declarative networking has been recently proposed to ease the task of programming in the cloud, by allowing the programmer to express only the desired result and leave the implementation details to the responsibility of the run-time system. In this context, we propose a study on a logic-programming-based computational model for eventually consistent, data-parallel systems, the keystone of which is provided by the recent finding that the class of programs that can be computed in an eventually consistent, coordination-free way is that of monotonic programs. This principle is called CALM and has been proven by Ameloot et al. for distributed, asynchronous settings. We advocate that CALM should be employed as a basic theoretical tool also for data-parallel systems, wherein computation usually proceeds synchronously in rounds and where communication is assumed to be reliable. It is general opinion that coordination-freedom can be seen as a major discriminant factor. In this work we make the case that the current form of CALM does not hold in general for data-parallel systems, and show how, using novel techniques, the satisfiability of the CALM principle can still be obtained although just for the subclass of programs called connected monotonic queries. We complete the study with considerations on the relationships between our model and the one employed by Ameloot et al., showing that our techniques subsume the latter when the synchronization constraints imposed on the system are loosened.
Heterogeneous systems are becoming more common on High Performance Computing (HPC) systems. Even using tools like CUDA and OpenCL it is a non-trivial task to obtain optimal performance on the GPU. Approaches to simplifying this task include Merge (a library based framework for heterogeneous multi-core systems), Zippy (a framework for parallel execution of codes on multiple GPUs), BSGP (a new programming language for general purpose computation on the GPU) and CUDA-lite (an enhancement to CUDA that transforms code based on annotations). In addition, efforts are underway to improve compiler tools for automatic parallelization and optimization of affine loop nests for GPUs and for automatic translation of OpenMP parallelized codes to CUDA. In this paper we present an alternative approach: a new computational framework for the development of massively data parallel scientific codes applications suitable for use on such petascale/exascale hybrid systems built upon the highly scalable Cactus framework. As the first non-trivial demonstration of its usefulness, we successfully developed a new 3D CFD code that achieves improved performance.
The data-driven computing paradigm initially introduced by Kirchdoerfer and Ortiz (2016) enables finite element computations in solid mechanics to be performed directly from material data sets, without an explicit material model. From a computational effort point of view, the most challenging task is the projection of admissible states at material points onto their closest states in the material data set. In this study, we compare and develop several possible data structures for solving the nearest-neighbor problem. We show that approximate nearest-neighbor (ANN) algorithms can accelerate material data searches by several orders of magnitude relative to exact searching algorithms. The approximations are suggested by--and adapted to--the structure of the data-driven iterative solver and result in no significant loss of solution accuracy. We assess the performance of the ANN algorithm with respect to material data set size with the aid of a 3D elasticity test case. We show that computations on a single processor with up to one billion material data points are feasible within a few seconds execution time with a speedup of more than 106 with respect to exact k-d trees.
Computational science is changing to be data intensive. Super-Computers must be balanced systems; not just CPU farms but also petascale IO and networking arrays. Anyone building CyberInfrastructure should allocate resources to support a balanced Tier-1 through Tier-3 design.
It is important for big data systems to identify their performance bottleneck. However, the popular indicators such as resource utilizations, are often misleading and incomparable with each other. In this paper, a novel indicator framework which can directly compare the impact of different indicators with each other is proposed to identify and analyze the performance bottleneck efficiently. A methodology which can construct the indicator from the performance change with the CPU frequency scaling is described. Spark is used as an example of a big data system and two typical SQL benchmarks are used as the workloads to evaluate the proposed method. Experimental results show that the proposed method is accurate compared with the resource utilization method and easy to implement compared with some white-box method. Meanwhile, the analysis with our indicators lead to some interesting findings and valuable performance optimization suggestions for big data systems.
A distributed system keeps consistency by disallowing data anomalies. However, especially in the database, the definitions of data anomalies in the current ANSI standard are controversial. The standard does not include all anomalies and does not introduce characters of anomalies. First, the definitions lack a mathematical formalization and cause ambiguous interpretations. Second, the definitions of anomalies are case-by-case, which could not have a comprehensive understanding of data anomalies. In this paper, we propose a ring anomalies detection method (the bingo model) in the distribution system and applying it to databases. The bingo model introduces anomalies construction and gives the base anomalies formalization method. Based on anomalies we propose consistency levels. We prove the simplified anomaly rings in the model to classified anomalies to give the independent consistency levels. We specify the bingo model to databases and find 22 anomalies in addition to existing anomalies.