Column-store in-memory databases have received a lot of attention because of their fast query processing response times on modern multi-core machines. Among different database operations, group by/aggregate is an impo...
详细信息
ISBN:
(纸本)9781467376846
Column-store in-memory databases have received a lot of attention because of their fast query processing response times on modern multi-core machines. Among different database operations, group by/aggregate is an important and potentially costly operation. Moreover, sort-based and hash-based algorithms are the most common ways of processing group by/aggregate queries. While sort-based algorithms are used in traditional DataBase Management systems(DBMS), hash based algorithms can be applied for faster query processing in new columnar databases. Besides, Graphical Processing Units(GPU) can be utilized as fast, high bandwidth co-processors to improve the query processing performance of columnar databases. The focus of this article is on the prototype for group by/aggregate operations that we created to exploit GPUs. We show different hash based algorithms to improve the performance of group by/aggregate operations on GPU. One of the parameters that affect the performance of the group by/aggregate algorithm is the number of groups and hashing algorithm. We show that we can get up to 7.6x improvement in kernel performance compared to a multi-core CPU implementation when we use a partitioned multilevel hash algorithm using GPU shared and global memories.
Earth science (ES) applications handle very large geospatial data sets and interactive response time is required by its query processing. Spatial selection is one of the very important basic operations for gee-spatial...
详细信息
ISBN:
(纸本)0769511538
Earth science (ES) applications handle very large geospatial data sets and interactive response time is required by its query processing. Spatial selection is one of the very important basic operations for gee-spatial databases. It retrieves all the objects that intersect with a given point or rectangle. This paper presents a novel approach for the parallel processing of spatial selection of very large geospatial databases using partitioned parallelism. To evaluate this approach, we use the Extended Sequoia 2000 benchmark, which has real world data and real queries. In addition, we use an actual object database management system, ShusseUo, which we developed previously. The experimental results of parallel processing of spatial selection show good speed-up.
We report a comparative study on the development of high-performance software for solving the problem of optimal design of multiproduct batch plants on modern parallelsystems. We analyze two main algorithmic approach...
详细信息
ISBN:
(纸本)9781538655559
We report a comparative study on the development of high-performance software for solving the problem of optimal design of multiproduct batch plants on modern parallelsystems. We analyze two main algorithmic approaches to optimization - branch-and-bound and metaheuristic-based - and we develop and compare their parallel implementations on a variety of parallel architectures: multi-core CPU, GPU, and clusters. Our experiments on a real-world case study - optimization of chemical-engineering systems - demonstrate the trade-offs between the run time performance and the quality of solutions achieved by different algorithms on various parallel architectures.
Error control codes have been widely adopted to design dependable storage systems, thus number of data accesses becomes numerous. Data access efficiency as well as energy saving are important problems. This paper prop...
详细信息
ISBN:
(纸本)9781479955848
Error control codes have been widely adopted to design dependable storage systems, thus number of data accesses becomes numerous. Data access efficiency as well as energy saving are important problems. This paper proposes a method which uses variable information word length for improving data access efficiency in dependable storage systems.
Graph-based structures are being increasingly used to model data and relations among data in a number of fields. Graph-based databases are becoming more popular as a means to better represent such data. Graph traversa...
详细信息
ISBN:
(纸本)9780769546759
Graph-based structures are being increasingly used to model data and relations among data in a number of fields. Graph-based databases are becoming more popular as a means to better represent such data. Graph traversal is a key component in graph algorithms such as reachability and graph matching. Since the scale of data stored and queried in these databases is increasing, it is important to obtain high performing implementations of graph traversal that can efficiently utilize the processing power of modern processors. In this work, we present a scalable Breadth-First Search Traversal algorithm for modern multi-socket, multi-core CPUs. Our algorithm uses lock- and atomic-free operations on a cache-resident structure for arbitrary sized graphs to filter out expensive main memory accesses, and completely and efficiently utilizes all available bandwidth resources. We propose a work distribution approach for multi-socket platforms that ensures load-balancing while keeping cross-socket communication low. We provide a detailed analytical model that accurately projects the performance of our single- and multi-socket traversal algorithms to within 5-10% of obtained performance. Our analytical model serves as a useful tool to analyze performance bottlenecks on modern CPUs. When measured on various synthetic and real-world graphs with a wide range of graph sizes, vertex degrees and graph diameters, our implementation on a dual-socket Intel (R) Xeon (R) X5570 (Intel microarchitecture code name Nehalem) system achieves 1.5X-13.2X performance speedup over the best reported numbers. We achieve around 1 Billion traversed edges per second on a scale-free R-MAT graph with 64M vertices and 2 Billion edges on a dual-socket Nehalem system. Our optimized algorithm is useful as a building block for efficient multi-node implementations and future exascale systems, thereby allowing them to ride the trend of increasing per-node compute and bandwidth resources.
This paper reviews the problem of correctness in the context of scaleable distributedsystems. It takes the concept of a strongly eventually consistent (SEC) data structure based on the work done on conflict free repl...
详细信息
ISBN:
(纸本)9781728189468
This paper reviews the problem of correctness in the context of scaleable distributedsystems. It takes the concept of a strongly eventually consistent (SEC) data structure based on the work done on conflict free replicated data types and examines the theoretical basis of a practical application to generalised database management systems. It reasons that the condition required for a database management system to implement a monotonic semilattice and remain generic is not possible. It further proposes a framework which extends the SEC condition to capture enough information to make such a system practical to implement. Finally, it evaluates such an implementation using a modified storage engine for Apache Ignite demonstrating not just its practicality but its scaleability in Ignite and other database management systems like it. The key contributions of this work are Semantic and intent defined conflict free replicated data types, an enriched CRDT which enables generalised database systems to meet the SEC criteria for a monotonic semilattice object and Semantic Intent Configuration Language, SICL, a language and framework for distributed computing.
The PDSE'99 symposium, now in its fourth successive year, is a continuation of the inaugural PDSE'96 workshop held in conjunction with ICSE-18 in Berlin. This was followed by ICSE events in Boston, 1997 and in...
详细信息
ISBN:
(纸本)9781581130744
The PDSE'99 symposium, now in its fourth successive year, is a continuation of the inaugural PDSE'96 workshop held in conjunction with ICSE-18 in Berlin. This was followed by ICSE events in Boston, 1997 and in Kyoto, 1998. The PDSE symposiums attract up to 50 delegates from a wide range of backgrounds from both industry and academia from around the world. The papers are always published internationally, more recently as IEEE proceedings, and are refereed by a minimum of three independent experts in the field. The talks include keynote speakers with an emphasis on first hand experience of industrial applications and large-scale projects.
We consider closed pattern mining from distributed multi-relational databases, especially focusing on its efficient implementation. Given a set of local databases (horizontal partitions), we first compute their sets o...
详细信息
ISBN:
(纸本)9781479959556
We consider closed pattern mining from distributed multi-relational databases, especially focusing on its efficient implementation. Given a set of local databases (horizontal partitions), we first compute their sets of closed patterns (concepts) using a closed pattern mining algorithm tailored to multi-relational data mining (MRDM). We then generate the set of closed patterns in the global database by utilizing the merge (or subposition) operator, studied in the field of Formal Concept Analysis. Since the computational complexity of MRDM increases compared with the conventional itemset mining, we propose some methods for improving the overall computations. We also present some experimental results using a distributed computation environment based on the MapReduce framework, which shows the effectiveness of the proposed methods.
Digital images are being generated at a phenomenal rate. Currently, the most common method by far for searching digital image databases is based on employing index terms which are entered manually. Content based image...
详细信息
ISBN:
(纸本)1892512459
Digital images are being generated at a phenomenal rate. Currently, the most common method by far for searching digital image databases is based on employing index terms which are entered manually. Content based image retrieval (CM) systems are required to effectively and efficiently use the information that is intrinsically stored in these image databases. However, CBIR on multiple image databases is a slow process that needs performance improvement. In this paper, we present a distributed approach to CBIR, an image retrieval scheme that retrieves images based on segmentation and signature in a distributed environment.
In this paper, we present the use of parallel-in-space simulation approach to accelerate the dynamic simulation of power systems with high penetration of distributed generation, i. e. high number of power electronic d...
详细信息
ISBN:
(数字)9781665466189
ISBN:
(纸本)9781665466189
In this paper, we present the use of parallel-in-space simulation approach to accelerate the dynamic simulation of power systems with high penetration of distributed generation, i. e. high number of power electronic devices. The approach is implemented using the OpenCL framework and executed on a graphic processing unit (GPU). We benchmark our prototype implementation using a distribution network with increasing number of distributed generator which is modeled by voltage source inverter. Results show that the computation time for the distributed generators solution stays nearly constant with the increasing number of the distributed generators in the network.
暂无评论