Browsing by Subject "Iterative improvements"

Now showing 1 - 3 of 3

Open Access
A novel method for scaling iterative solvers: avoiding latency overhead of parallel sparse-matrix vector multiplies
(Institute of Electrical and Electronics Engineers, 2015) Selvitopi, R. O.; Ozdal, M. M.; Aykanat, Cevdet
In parallel linear iterative solvers, sparse matrix vector multiplication (SpMxV) incurs irregular point-to-point (P2P) communications, whereas inner product computations incur regular collective communications. These P2P communications cause an additional synchronization point with relatively high message latency costs due to small message sizes. In these solvers, each SpMxV is usually followed by an inner product computation that involves the output vector of SpMxV. Here, we exploit this property to propose a novel parallelization method that avoids the latency costs and synchronization overhead of P2P communications. Our method involves a computational and a communication rearrangement scheme. The computational rearrangement provides an alternative method for forming input vector of SpMxV and allows P2P and collective communications to be performed in a single phase. The communication rearrangement realizes this opportunity by embedding P2P communications into global collective communication operations. The proposed method grants a certain value on the maximum number of messages communicated regardless of the sparsity pattern of the matrix. The downside, however, is the increased message volume and the negligible redundant computation. We favor reducing the message latency costs at the expense of increasing message volume. Yet, we propose two iterative-improvement-based heuristics to alleviate the increase in the volume through one-to-one task-to-processor mapping. Our experiments on two supercomputers, Cray XE6 and IBM BlueGene/Q, up to 2,048 processors show that the proposed parallelization method exhibits superior scalable performance compared to the conventional parallelization method.
Open Access
Query-log aware replicated declustering
(Institute of Electrical and Electronics Engineers, 2013) Turk, A.; Oktay, K. Y.; Aykanat, Cevdet
Data declustering and replication can be used to reduce I/O times related with processing of data intensive queries. Declustering parallelizes the query retrieval process by distributing the data items requested by queries among several disks. Replication enables alternative disk choices for individual disk items and thus provides better query parallelism options. In general, existing replicated declustering schemes do not consider query log information and try to optimize all possible queries for a specific query type, such as range or spatial queries. In such schemes, it is assumed that two or more copies of all data items are to be generated and scheduling of these copies to disks are discussed. However, in some applications, generation of even two copies of all of the data items is not feasible, since data items tend to have very large sizes. In this work, we assume that there is a given limit on disk capacities and thus on replication amounts. We utilize existing query-log information to propose a selective replicated declustering scheme, in which we select the data items to be replicated and decide on their scheduling onto disks while respecting disk capacities. We propose and implement an iterative improvement algorithm to obtain a two-way replicated declustering and use this algorithm in a recursive framework to generate a multiway replicated declustering. Then we improve the obtained multiway replicated declustering by efficient refinement heuristics. Experiments conducted on realistic data sets show that the proposed scheme yields better performance results compared to existing replicated declustering schemes. © 1990-2012 IEEE.
Open Access
Selective replicated declustering for arbitrary queries
(Springer, 2009-08) Oktay, K. Yasin; Türk, Ata; Aykanat, Cevdet
Data declustering is used to minimize query response times in data intensive applications. In this technique, query retrieval process is parallelized by distributing the data among several disks and it is useful in applications such as geographic information systems that access huge amounts of data. Declustering with replication is an extension of declustering with possible data replicas in the system. Many replicated declustering schemes have been proposed. Most of these schemes generate two or more copies of all data items. However, some applications have very large data sizes and even having two copies of all data items may not be feasible. In such systems selective replication is a necessity. Furthermore, existing replication schemes are not designed to utilize query distribution information if such information is available. In this study we propose a replicated declustering scheme that decides both on the data items to be replicated and the assignment of all data items to disks when there is limited replication capacity. We make use of available query information in order to decide replication and partitioning of the data and try to optimize aggregate parallel response time. We propose and implement a Fiduccia-Mattheyses-like iterative improvement algorithm to obtain a two-way replicated declustering and use this algorithm in a recursive framework to generate a multi-way replicated declustering. Experiments conducted with arbitrary queries on real datasets show that, especially for low replication constraints, the proposed scheme yields better performance results compared to existing replicated declustering schemes. © 2009 Springer.