Indexing of Spatiotemporal Trajectories for Efficient Distance Threshold Similarity Searches on the GPU

Gowanlock, Michael; Casanova, Henri

doi:10.1109/ipdps.2015.24

Cited by 13 publications

(10 citation statements)

References 29 publications

(40 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More recently, grid-based similarity measures that yield relatively low complexity have become popular [14]. Still, the computational intensity of implementing the similarity query on massive trajectory data often becomes a bottleneck [8,15]. Our goal is not to formulate a new similarity measure, but to design parallel solutions to support the execution of the query using the K-BCT query as an example.…”

Section: Computational Solutions To Similarity Queriesmentioning

confidence: 99%

“…A MBR can be generated from any geometric feature such as a polyline or a multi-part polygon. Scientists have employed MBRs as a filtering method in the nearest neighbor (NN) search which identifies the nearest feature to a given feature [8]. MBR-based filtering process usually calculates the distances between the MBRs of features [17] (e.g., complex polygons) and sorts the features based on the distances.…”

Section: Mbr and Mindist For Efficient Searchingmentioning

confidence: 99%

“…Besides providing measures for sorting trajectories, MINDIST can be used to terminate the search at an earlier stage. In the case of a NN search, when the distance between a pair of objects is smaller than the MINDIST of the rest objects, the pair of NN objects is found [8]. Therefore, we can use MINDIST to speed up the SD computation.…”

Section: Mbr and Mindist For Efficient Searchingmentioning

confidence: 99%

“…Algorithm 1 is first applied to the trajectories in Phase 1 and then applied to the trajectories in Phase 2. To identify the shortest distance and the matched point for every query point, we introduce a parallel simple grid searching method [8] to speed up the shortest distance identification [14,15]. This idea is based on a few grid-based similarity measures (e.g., [20]).…”

Section: Definitionmentioning

confidence: 99%

See 3 more Smart Citations

Efficient Parallel K Best Connected Trajectory (K-BCT) Query with GPGPU: A Combinatorial Min-Distance and Progressive Bounding Box Approach

Wang

Zhang

et al. 2018

IJGI

View full text Add to dashboard Cite

Performing similarity analysis on trajectories consisting of massive numbers of tracking points is computationally challenging. We introduce a progressive minimum bounding rectangle (MBR) and minimum distance (MINDIST) approach to process the K Best Connected Trajectory (K-BCT) query, which aims to find the top K similarity trajectories to a given query trajectory. Our approach has three unique features to speed up the query. First, the approach builds a series of progressive MBRs from the query trajectory to determine the order of reference trajectories to identify the target top K reference trajectories at an earlier stage. Second, this method introduces a grid-based search method to speed up the matched point detection between two trajectories for similarity measures. Third, this approach further leverages the many-core computing power of Graphical Processing Unit (GPU) devices to perform the query in a parallel manner. We have conducted tests with ship tracking data and human movement data using GPU instances from Amazon Web Services. Preliminary results indicate that (a) parallel computing has greatly improved the efficiency of the query, and (b) our optimized approach can further speedup the computation compared to parallel implementations.

show abstract

Section: Computational Solutions To Similarity Queriesmentioning

confidence: 99%

Section: Mbr and Mindist For Efficient Searchingmentioning

confidence: 99%

Section: Mbr and Mindist For Efficient Searchingmentioning

confidence: 99%

Section: Definitionmentioning

confidence: 99%

See 2 more Smart Citations

Efficient Parallel K Best Connected Trajectory (K-BCT) Query with GPGPU: A Combinatorial Min-Distance and Progressive Bounding Box Approach

Wang

Zhang

et al. 2018

IJGI

View full text Add to dashboard Cite

show abstract

“…ere are two major indexing strategies for the GPU: (i) index-trees, similar to those that have been shown to provide good performance on the CPU, such as the R-tree [24]; or (ii) non-hierarchical indexes, such as grids or binning. Several works propose e cient indexes for points or other objects on the GPU [12,20,21,28,29,38].Kim et al [28] designed an R-tree for the GPU to optimize index searches that avoids many of the drawbacks of executing tree traversals on the GPU. Later, the same research group presented a hybrid approach [29] that splits the R-tree between the CPU and GPU by assigning parts of the algorithm with more regular and irregular instruction ows to the GPU and CPU, respectively.…”

mentioning

confidence: 99%

GPU-Accelerated Similarity Self-Join for Multi-Dimensional Data

Gowanlock

Karsin

2019

Proceedings of the 15th International Workshop on Data Management on New Hardware

Self Cite

View full text Add to dashboard Cite

e self-join nds all objects in a dataset that are within a search distance, ϵ , of each other; therefore, the self-join is a building block of many algorithms. We advance a GPU-accelerated self-join algorithm targeted towards high dimensional data. e massive parallelism a orded by the GPU and high aggregate memory bandwidth makes the architecture well-suited for data-intensive workloads. We leverage a grid-based, GPU-tailored index to perform range queries. We propose the following optimizations: (i) a trade-o between candidate set ltering and index search overhead by exploiting properties of the index; (ii) reordering the data based on variance in each dimension to improve the ltering power of the index; and (iii) a pruning method for reducing the number of expensive distance calculations. Across most scenarios on real-world and synthetic datasets, our algorithm outperforms the parallel state-of-the-art approach. Exascale systems are converging on heterogeneous distributed-memory architectures. We show that an entity partitioning method can be utilized to achieve a balanced workload, and thus good scalability for multi-GPU or distributed-memory self-joins.us, many large-scale data analytics applications will rely on GPU-e cient algorithms, including the distance similarity self-join for high dimensional data -the subject of this work. is paper makes the following novel contributions:• Leveraging an e cient indexing scheme for the GPU, we exploit the trade-o between index ltering power and search cost to improve the overall performance of searching high dimensional feature spaces.• We improve the ltering power of the index by reordering the data in each dimension using statistical properties of the data distribution. We show that this is particularly important when exploiting the trade-o outlined above.• We mitigate the performance cost of reducing index ltering power by proposing a technique that prunes the candidate set by comparing points based on an un-indexed dimension.• We show that on the worst-case data distribution for our approach, we achieve signi cantly be er performance than the state-of-the-art on the same scenario. is suggests that the performance of the GPU-accelerated self-join is resilient to the data distribution, making the approach well-suited for many application scenarios.• We evaluate our approach on 5 real-world and 3 synthetic datasets and show that our GPU accelerated self-join outperforms the state-of-the-art parallel algorithm in the literature.• e self-join is an expensive operation. We show initial insights into the scalability of the self-join on multi-GPU and distributed-memory systems, and demonstrate that an entity partitioning strategy can be used to achieve good load balancing. e paper is outlined as follows: Section 2 provides background material, Section 3 formalizes the problem and discusses previous work that we employ, Section 4 presents the novel methods we use to improve high dimensional self-join performance, Section 5 illustrates our performance results, Section 6 dis...

show abstract