Subhajit Sahu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Subhajit Sahu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
Kishore Kothapalli
IIIT Hyderabad,Hyderabad,India,500032
12 shared publicationsDip Sankar Banerjee
IIT Jodhpur,Karwar,India,342030
9 shared publications- Kishore Kothapalli
International Institute of Information Technology Hyderabad, Hyderabad, India
7 shared publications - Dip Sankar Banerjee
Indian Institute of Technology Jodhpur, Jodhpur, Rajasthan, INDIA
3 shared publications Hemalatha Eedi
JNTUH College of Engineering,Hyderabad,India
1 shared publication- Hemalatha Eedi
JNTUH College of Engineering Hyderabad
1 shared publication
Latest Publications
- Open Access
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 509-520
νMG-LPA and νBM-LPA: Memory Efficient GPU-based Label Propagation Algorithms (LPA) for Community Detection
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 509-520
Open AccessνMG-LPA and νBM-LPA: Memory Efficient GPU-based Label Propagation Algorithms (LPA) for Community Detection
- Open Access
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 498-508
Efficient Tracking of Communities on Evolving Graphs with Leiden Algorithm
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 498-508
Open AccessEfficient Tracking of Communities on Evolving Graphs with Leiden Algorithm
- Open Access
- 16 April 2026
- Published by MDPI AG in Future Internet
- Vol. 18 (4) , 212
Community-Aware Network Dismantling via Gateways: Large-Scale Evaluation on LFR Benchmarks
Network dismantling—the targeted removal of nodes to degrade large-scale connectivity—plays a central role in resilience analysis, epidemic containment, and systemic-risk mitigation. Recent work shows that dismantling performance depends strongly on mesoscale modular structure, suggesting that community-aware strategies may offer advantages over classical centrality-based heuristics. In this work, we perform a large-scale, systematic evaluation of dismantling strategies and introduce gateways as a new mesoscale dismantling concept. While similar experiments exist using degree- and betweenness-based dismantling strategies, we check a new strategy based on gateways, which capture asymmetric entry points into communities and generalize the notion of inter-community connectors. Furthermore, we process a massive dataset of 568,584 LFR benchmark graphs, covering a wide range of degree distributions, community sizes, and mixing parameters. For evaluation, we use both extrinsic (ARI, NMI, FMI, VI) and intrinsic (Modularity, Coverage, Performance, Average Conductance, Average Internal Density) metrics. We find that across parameter regimes and evaluation metrics, classical strategies (degree, betweenness, community connections) and gateway-based dismantling exhibit broadly similar performance. Our results also corroborate recent findings that dismantling effectiveness is robust to the specific partitioning algorithm and that inter-community connectivity plays a dominant role in global fragmentation. The evaluation provides large-scale evidence that gateway-aware dismantling captures an operationally relevant mesoscale mechanism as good as previous approaches and motivates further empirical studies on real networks and cost-aware settings.
- 16 April 2026
- Published by MDPI AG in Future Internet
- Vol. 18 (4) , 212
Open AccessCommunity-Aware Network Dismantling via Gateways: Large-Scale Evaluation on LFR Benchmarks
Network dismantling—the targeted removal of nodes to degrade large-scale connectivity—plays a central role in resilience analysis, epidemic containment, and systemic-risk mitigation. Recent work shows that dismantling performance depends strongly on mesoscale modular structure, suggesting that community-aware strategies may offer advantages over classical centrality-based heuristics. In this work, we perform a large-scale, systematic evaluation of dismantling strategies and introduce gateways as a new mesoscale dismantling concept. While similar experiments exist using degree- and betweenness-based dismantling strategies, we check a new strategy based on gateways, which capture asymmetric entry points into communities and generalize the notion of inter-community connectors. Furthermore, we process a massive dataset of 568,584 LFR benchmark graphs, covering a wide range of degree distributions, community sizes, and mixing parameters. For evaluation, we use both extrinsic (ARI, NMI, FMI, VI) and intrinsic (Modularity, Coverage, Performance, Average Conductance, Average Internal Density) metrics. We find that across parameter regimes and evaluation metrics, classical strategies (degree, betweenness, community connections) and gateway-based dismantling exhibit broadly similar performance. Our results also corroborate recent findings that dismantling effectiveness is robust to the specific partitioning algorithm and that inter-community connectivity plays a dominant role in global fragmentation. The evaluation provides large-scale evidence that gateway-aware dismantling captures an operationally relevant mesoscale mechanism as good as previous approaches and motivates further empirical studies on real networks and cost-aware settings.
- 9 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
Towards Dynamic Community Detection with Leiden Algorithm
Community detection is the process of grouping nodes in a network into clusters. However, real-world graphs often evolve over time, making repeated community detection on such graphs expensive. In this paper, we extend three dynamic approaches, namely, Naive-dynamic (ND), Delta-screening (DS), and Dynamic Frontier (DF), to a fast multicore implementation of the Leiden algorithm - an algorithm known for its high-quality community detection - using subset renumbering, selective refinement, and load balancing of the aggregation phase. This is, to the best of our knowledge, the first attempt at applying such dynamic approaches to the Leiden algorithm.
- 9 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
Towards Dynamic Community Detection with Leiden Algorithm
Community detection is the process of grouping nodes in a network into clusters. However, real-world graphs often evolve over time, making repeated community detection on such graphs expensive. In this paper, we extend three dynamic approaches, namely, Naive-dynamic (ND), Delta-screening (DS), and Dynamic Frontier (DF), to a fast multicore implementation of the Leiden algorithm - an algorithm known for its high-quality community detection - using subset renumbering, selective refinement, and load balancing of the aggregation phase. This is, to the best of our knowledge, the first attempt at applying such dynamic approaches to the Leiden algorithm.
- 9 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
GPU-Accelerated Dynamic Frontier PageRank
PageRank is a centrality measure that ranks graph vertices based on their connections and importance. This paper presents efficient GPU implementations for both Static PageRank and incrementally expanding (and contracting) Dynamic Frontier (DF) and Dynamic Frontier with Pruning (DFP) PageRank, which process only a subset of vertices likely to change ranks. The Static PageRank implementation recomputes scores from scratch using a synchronous, pull-based, atomicsfree approach. It partitions vertices into low and high in-degrees, processed by two separate kernels. Our DF and DF-P PageRank additionally partition low and high out-degree vertices, using two additional kernels for incremental expansion and/or contraction of the set of affected vertices. On an NVIDIA A100 GPU, our Static PageRank outperforms Hornet and Gunrock’s implementations by $16.3 \times$ and $5.3 \times$, respectively. Further, our DF and DF-P PageRank achieve speedups of $5.2 \times / 3.9 \times$ on real-world dynamic graphs and $12.8 \times / 6.2 \times$ on large static graphs with random updates compared to prior multicore implementations.- 9 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
GPU-Accelerated Dynamic Frontier PageRank
PageRank is a centrality measure that ranks graph vertices based on their connections and importance. This paper presents efficient GPU implementations for both Static PageRank and incrementally expanding (and contracting) Dynamic Frontier (DF) and Dynamic Frontier with Pruning (DFP) PageRank, which process only a subset of vertices likely to change ranks. The Static PageRank implementation recomputes scores from scratch using a synchronous, pull-based, atomicsfree approach. It partitions vertices into low and high in-degrees, processed by two separate kernels. Our DF and DF-P PageRank additionally partition low and high out-degree vertices, using two additional kernels for incremental expansion and/or contraction of the set of affected vertices. On an NVIDIA A100 GPU, our Static PageRank outperforms Hornet and Gunrock’s implementations by $16.3 \times$ and $5.3 \times$, respectively. Further, our DF and DF-P PageRank achieve speedups of $5.2 \times / 3.9 \times$ on real-world dynamic graphs and $12.8 \times / 6.2 \times$ on large static graphs with random updates compared to prior multicore implementations.- 15 July 2025
- Published by Elsevier in Future Generation Computer Systems
- Vol. 174
GVE-LPA and GSL-LPA: High-speed and internally-connected label propagation on multicore systems
- 15 July 2025
- Published by Elsevier in Future Generation Computer Systems
- Vol. 174
GVE-LPA and GSL-LPA: High-speed and internally-connected label propagation on multicore systems
- 11 June 2025
- Published by Springer Nature
- p. 268-280
GVEL: Fast Graph Loading in Edgelist and Compressed Sparse Row (CSR) Formats
- 11 June 2025
- Published by Springer Nature
- p. 268-280
GVEL: Fast Graph Loading in Edgelist and Compressed Sparse Row (CSR) Formats
- 10 June 2025
- Published by Springer Nature
- p. 361-366
High-Speed Neighborhood-Based Link Prediction by Disregarding Large Hubs
- 10 June 2025
- Published by Springer Nature
- p. 361-366
High-Speed Neighborhood-Based Link Prediction by Disregarding Large Hubs
- 10 June 2025
- Published by Springer Nature
- p. 93-104
GVE-LPA: Fast Label Propagation Algorithm (LPA) for Community Detection in the Shared Memory Setting
- 10 June 2025
- Published by Springer Nature
- p. 93-104
GVE-LPA: Fast Label Propagation Algorithm (LPA) for Community Detection in the Shared Memory Setting
- 3 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 395-404
ν-LPA: Fast GPU-based Label Propagation Algorithm (LPA) for Community Detection
Community detection is the problem of identifying natural divisions in networks. Efficient parallel algorithms for identifying such divisions are critical in a number of applications. This paper presents an optimized implementation of the Label Propagation Algorithm (LPA) for community detection, featuring an asynchronous LPA with a Pick-Less (PL) method every 4 iterations to handle community swaps, ideal for SIMT hardware like GPUs. It also introduces a novel per-vertex hashtable with hybrid quadratic-double probing for collision resolution. On an NVIDIA A100 GPU, our implementation, ν-LPA, outperforms FLPA (sequential), NetworKit LPA (multicore), Gunrock LPA (GPU), and cuGraph Louvain (GPU) by 364×, 62×, 2.6×, and 37×, respectively, while running FLPA and NetworKit LPA on a server with dual 16-core and achieves higher modularity than FLPA, but lower than NetworKit LPA and cuGraph Louvain.
- 3 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 395-404
ν-LPA: Fast GPU-based Label Propagation Algorithm (LPA) for Community Detection
Community detection is the problem of identifying natural divisions in networks. Efficient parallel algorithms for identifying such divisions are critical in a number of applications. This paper presents an optimized implementation of the Label Propagation Algorithm (LPA) for community detection, featuring an asynchronous LPA with a Pick-Less (PL) method every 4 iterations to handle community swaps, ideal for SIMT hardware like GPUs. It also introduces a novel per-vertex hashtable with hybrid quadratic-double probing for collision resolution. On an NVIDIA A100 GPU, our implementation, ν-LPA, outperforms FLPA (sequential), NetworKit LPA (multicore), Gunrock LPA (GPU), and cuGraph Louvain (GPU) by 364×, 62×, 2.6×, and 37×, respectively, while running FLPA and NetworKit LPA on a server with dual 16-core and achieves higher modularity than FLPA, but lower than NetworKit LPA and cuGraph Louvain.