Publications

Dr. Top-k: Delegate-Centric Top-k on GPUs

Published in Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC), 2021

Dr. Top-k, is a Delegate-centric top-k system on GPUs that can reduce the workloads of top-k methods significantly. In particular, the works contains three major contributions: First, we introduce a comprehensive design of the delegate-centric concept, including maximum delegate, delegate-based filtering, and β delegate mechanisms to help reduce the workload for top-k up to more than 99%. Second, due to the difficulty and importance of deriving a proper subrange size, we perform a rigorous theoretical analysis, coupled with thorough experimental validations to identify the desirable subrange size. Third, we introduce four key system optimizations to enable fast multi-GPU top-k computation. Taken together, this work constantly outperforms the state-of-the-art.

Recommended citation: Gaihre, Anil, et al. "Dr.Top-k: Delegate-Centric Top-k Computation on GPUs." In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC), 2021. https://anil-gaihre.github.io/files/DrTopk.pdf

GSOFA: Scalable Sparse Symbolic LU Factorization on GPUs

Published in IEEE TPDS Special Section: Innovative R&D toward the Exascale Era, 2021

This paper is about GPU based symbolic factorization for LU decomposition on sparse direct linear solvers.

Recommended citation: Gaihre, Anil, Xiaoye S. Li, and Hang Liu. "GSOFA: Scalable Sparse LU Symbolic Factorization on GPUs." arXiv preprint arXiv:2007.00840 (2020) https://anil-gaihre.github.io/files/GSOFA.pdf

XBFS: eXploring Runtime Optimizations for Breadth-First Search on GPUs

Published in HPDC19: Proceedings of the 28th International Symposium on High-Performance Parallel and Distributed Computing, 2019

XBFS introduces dynamic optimizations to BFS on GPUs. It adaptively uses four either novel or optimized scan approaches to rapidly generate frontier queue. Further, inspired by the observation that bottom-up BFS experiences unpredictable amounts of workload, the paper proposes the novel dynamic workload balancing method. Third, the work designs and implements the first truly asynchronous BFS traversal.

Recommended citation: Gaihre, Anil, et al. "Xbfs: exploring runtime optimizations for breadth-first search on gpus." Proceedings of the 28th International Symposium on High-Performance Parallel and Distributed Computing. 2019. https://anil-gaihre.github.io/files/XBFS.pdf

Deanonymizing Cryptocurrency With Graph Learning: The Promises and Challenges

Published in 2019 IEEE Conference on Communications and Network Security (CNS), 2019

This paper discusses the promises and challenges of exploiting graph learning to deanonymizing cryptocurrencies, which can aid the cyberfighters to circumvent cryptocurrency-based illicit activities.

Recommended citation: Gaihre, Anil, Santosh Pandey, and Hang Liu. "Deanonymizing cryptocurrency with graph learning: the promises and challenges." 2019 IEEE Conference on Communications and Network Security (CNS). IEEE, 2019. https://anil-gaihre.github.io/files/anonymity.pdf

Do bitcoins users really care about anonymity? An Analysis of the Bitcoin Transaction Graph

Published in 2018 IEEE International Conference on Big Data (Big Data), 2018

This paper examines the Bitcoin transaction graphs to answer two critical yet unanswered questions concerning anonymity and privacy: Do typical Bitcoin users care about anonymity? Do critical users care about anonymity?

Recommended citation: Gaihre, Anil, Yan Luo, and Hang Liu. "Do bitcoin users really care about anonymity? an analysis of the bitcoin transaction graph." 2018 IEEE International Conference on Big Data (Big Data). IEEE, 2018. https://anil-gaihre.github.io/files/anonymity.pdf

Anil Gaihre