CaltechAUTHORS
  A Caltech Library Service

Bigger Buffer k-d Trees on Multi-Many-Core Systems

Gieseke, Fabian and Oancea, Cosmin Eugen and Mahabal, Ashish and Igel, Christian and Heskes, Tom (2019) Bigger Buffer k-d Trees on Multi-Many-Core Systems. In: High Performance Computing for Computational Science. Lecture Notes in Computer Science. No.11333. Springer , Cham, Switzerland, pp. 202-214. https://resolver.caltech.edu/CaltechAUTHORS:20190325-133557161

[img] PDF - Submitted Version
See Usage Policy.

522Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20190325-133557161

Abstract

A buffer k-d tree is a k-d tree variant for massively-parallel nearest neighbor search. While providing valuable speed-ups on modern many-core devices in case both a large number of reference and query points are given, buffer k-d trees are limited by the amount of points that can fit on a single device. In this work, we show how to modify the original data structure and the associated workflow to make the overall approach capable of dealing with massive data sets. We further provide a simple yet efficient way of using multiple devices given in a single workstation. The applicability of the modified framework is demonstrated in the context of astronomy, a field that is faced with huge amounts of data.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
https://doi.org/10.1007/978-3-030-15996-2_15DOIArticle
https://rdcu.be/bJjyPPublisherFree ReadCube access
https://arxiv.org/abs/1512.02831arXivDiscussion Paper
ORCID:
AuthorORCID
Mahabal, Ashish0000-0003-2242-0244
Additional Information:© 2019 Springer Nature Switzerland AG. First Online: 26 March 2019. The authors would like to thank the Radboud Excellence Initiative of the Radboud University Nijmegen (FG), NVIDIA for generous hardware donations (FG), the Danish Industry Foundation through the Industrial Data Analysis Service (FG, CI, CO), the The Danish Council for Independent Research | Natural Sciences through the project Surveying the sky using machine learning (CI), and ACP, IUCAA, IUSSTF, and NSF (AM).
Funders:
Funding AgencyGrant Number
Danish Industry FoundationUNSPECIFIED
Danish Council for Independent ResearchUNSPECIFIED
ACP Observatory Control softwareUNSPECIFIED
Inter-University Centre for Astronomy and AstrophysicsUNSPECIFIED
Indo-U.S. Science and Technology ForumUNSPECIFIED
NSFUNSPECIFIED
Series Name:Lecture Notes in Computer Science
Issue or Number:11333
Record Number:CaltechAUTHORS:20190325-133557161
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20190325-133557161
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:94124
Collection:CaltechAUTHORS
Deposited By: Tony Diaz
Deposited On:25 Mar 2019 20:58
Last Modified:25 Nov 2019 17:51

Repository Staff Only: item control page