Pith. sign in

REVIEW 3 major objections 3 minor 41 references

GPU-Accelerated Multilevel Graph Clustering: A Parallel Perspective on Louvain and Leiden

T0 review · 3 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A GPU implementation of Leiden preserves all six sequential quality guarantees, and a GPU Louvain variant reports 3.1x–8.8x speedups with top modularity on 57 graphs.

desk verdict Strong empirical GPU-clustering paper whose central 'provable guarantees' claim rests on a faulty proof sketch; referees should check Lemma 2 before believing Theorem 1. read the letter →

arxiv 2608.01503 v1 pith:OXDW4ASQ submitted 2026-08-02 cs.DC

Michael S. Gilbert , Kamesh Madduri This is my paper · ORCID
classification cs.DC
keywords GPUgraphclusteringLouvainalgorithmLeidenmodularityLambdaCCmultilevelpartitioningparallelcommunitydetectionspanning-treerefinement
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that the two most widely used multilevel graph-clustering heuristics, Louvain and Leiden, can be parallelized on a GPU without sacrificing their quality guarantees. Its Leiden variant, pLeiden, is claimed to be the first parallel implementation that provably preserves all six guarantees of sequential Leiden, because its spanning-tree refinement can generate exactly the same clusterings as Leiden's original refinement. Its Louvain variant, pLouvain, adapts Louvain+ with a synchronous local-move scheme and a symmetry-breaking filter, and adds an uncoarsening iteration that fixes weak connectivity inside clusters. On 57 graphs, the paper reports geometric-mean speedups of 3.1x and 8.8x over the fastest open-source parallelizations, with pLouvain producing the highest modularity on nearly all graphs. If these claims hold, GPU implementations can match the quality of established sequential heuristics while running an order of magnitude faster.

What carries the argument

pLeidenR is the load-bearing object: each vertex in a constraint cluster selects a random incident edge (or self-loop) with non-negative LambdaCC weight $w'(u,v)\ge 0$, forming a random forest; a random ordering orients edges downward, subsampling to remove cycles, and each tree is sorted so every prefix is a connected vertex set. A commit-well-connected-prefixes step checks the two prefix-sum conditions LeidenR checks, so the set of reachable clusterings is identical to sequential LeidenR's (Lemmas 1 and 2). For pLouvain, the equivalent mechanism is the afterburner filter, a parallel symmetry-breaking pass that admits a candidate vertex move when the combined objective delta with previously

What would settle it

On a small graph, exhaustively enumerate the clusterings reachable by sequential LeidenR and by pLeidenR over all random seeds; any clustering reachable by only one side would falsify Lemma 1 or Lemma 2. Alternatively, run pLeiden to a stable iteration on any graph and check the six properties directly, for example node optimality, where a single vertex whose current cluster scores worse than another available cluster would falsify Theorem 1.

Watch

Extended reading notes

Core claim

pLeiden is the first parallel implementation claimed to provably preserve every quality guarantee of sequential Leiden: $\gamma$-separation, $\gamma$-connectivity, subpartition $\gamma$-density, node optimality, uniform $\gamma$-density, and subset optimality, all restated in the LambdaCC objective that generalizes modularity and the Constant Potts Model. The transfer of guarantees is carried by pLeidenR, a refinement that builds a random spanning forest within constraint clusters, orients and serializes trees by a random ordering, and commits only connected prefixes that pass the same non-negative-weight tests LeidenR applies. Lemmas 1 and 2 establish that pLeidenR can generate precisely th

Load-bearing premise

That pLeidenR's spanning-tree construction generates exactly the clusterings the sequential Leiden refinement can generate, and that the fallback local move's maximizing prefix is non-empty whenever any non-node-optimal vertex exists; if either fails, Theorem 1's transfer of all six Leiden guarantees to pLeiden collapses.

Editorial extensions

If this is right

  • pLeiden can be used in GPU pipelines with the same six guarantees as sequential Leiden, so downstream analyses that depend on $\gamma$-connectivity or node optimality can trust GPU results.
  • pLouvain's uncoarsening iteration scheme guarantees connected clusters at stable iterations, removing the main quality objection to iterating Louvain.
  • pLouvain and pLeiden+ reach modularity comparable to a leading memetic algorithm within ten GPU iterations, suggesting memetic search on top of GPU clustering is feasible.
  • The reported speedups (3.1x for pLouvain, 8.8x for pLeiden, at least 15x for graph contraction) are measured against the fastest open-source parallelizations; if reproduced, they make GPU the default target for large Louvain/Leiden workloads.
  • The afterburner filter, kernel fission, and gather-free hash-table contraction are separable subroutines that other Louvain-style parallelizations could adopt.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: the spanning-tree equivalence is a general recipe: any cluster-joining refinement whose reachable clusterings are prefix-closed can be parallelized with random forests and the same if-and-only-if proof, potentially transferring guarantees to objectives beyond LambdaCC.
  • Inference: the runtime claims are architecture-specific; a natural test is whether the 3.1x and 8.8x geometric-mean speedups persist on smaller GPUs, AMD GPUs, or multi-GPU settings, since only the correctness theorem is hardware-independent.
  • Inference: pLouvain's quality gain comes mainly from uncoarsening; extending pLeiden's guarantee machinery to allow uncoarsening passes outside the guaranteed iterations, as pLeiden+ already does, could combine pLouvain's modularity with pLeiden's stable-iteration guarantees.
  • Inference: the afterburner filter's epsilon tie-break and phi temperature schedule define a quality/speed frontier that the paper samples at only two settings; sweeping epsilon and phi would show how much modularity can be bought per unit of runtime.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The paper proposes pLouvain and pLeiden, GPU-accelerated implementations of the Louvain+ and Leiden multilevel clustering algorithms. pLouvain uses a synchronous local move with an afterburner symmetry-breaking filter, kernel fission, and a hash-based gather-free contraction; pLeiden adds a spanning-tree-based refinement scheme (pLeidenR) and claims, in Theorem 1, to provably preserve all six quality guarantees of sequential Leiden. The authors report geometric mean speedups of 3.1x and 8.8x over state-of-the-art open-source parallel Louvain/Leiden implementations on a 57-graph benchmark, together with an ablation study and artifacts.

Significance. The empirical evaluation is thorough: 57 graphs, 21 runs per configuration, median runtimes, standard deviations, fixed competitor commits, an ablation study, and reproducible artifacts are provided. If the guarantee theorem were correct, the paper would be the first parallel Leiden implementation with provable quality guarantees, which is a significant result. The restatement of the six Leiden guarantees in LambdaCC notation is also a useful contribution. However, the central theoretical proof has a serious gap, so the main claim is not currently established.

major comments (3)
  1. [§V-B, Lemma 2 and Theorem 1] The proof of Lemma 2 is invalid and the lemma is false as stated. The assertion that when LeidenR moves v to cluster X there must exist u∈X with w'(u,v)≥0 does not follow from the move's non-negative objective delta: LeidenR's condition is aggregate (w'(v,X) - w'(v,C[v]) ≥ 0), and for modularity-type weights w'(v,C[v]) is negative, so the aggregate can be non-negative while every individual w'(v,u) is negative. The proof also contains an impossible case ('or will not be in a singleton-cluster when it is visited'), since a vertex is non-singleton only after being visited. Finally, the instruction to set H[v]=(u,u) assigns a self-loop with w'(u,u)=-λw(u)^2<0, which Algorithm 6 cannot select because A_u requires w'(u,v)≥0. Since Theorem 1 relies on the if-and-only-if equivalence, and the asymptotic uniform γ-density and subset optimality guarantees require reachability of the relevant clust
  2. [§V-B, Algorithm 6] COMMITWELLCONNECTEDPREFIXES is invoked on line 9 but is never defined or specified. Both Lemma 1 and Lemma 2 depend on its exact semantics (which prefixes are committed, how the suffix is handled, and what conditions are checked). Without a formal specification or pseudocode, the proof cannot be verified. This should be added and the lemmas proved against that specification.
  3. [§IV-D] The paper claims that the new Louvain+ iteration method 'guarantees connected clusters in stable iterations' and lists this as contribution C6, but no proof is provided. This is a standalone correctness claim, not a mere implementation note, and needs either a proof or a reference to where it is proved.
minor comments (3)
  1. [§III-A] Typo: 'via a a multilevel algorithm' should be 'via a multilevel algorithm'.
  2. [§IV-A, Algorithm 2] The data structure DS is introduced in lines 2-5 but its concrete representation and update semantics are described only informally in the text. A short definition would improve reproducibility.
  3. [§VI-C] The pass limit and temperature schedule (φ=0.75 for four passes, φ=0.25 for two) are fixed; the ablation study only tests φ=0. A brief sensitivity discussion would strengthen the claim that these are not overfit to the benchmark set.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; Theorem 1 is supported by a constructive equivalence proof rather than by fitting or self-citation.

full rationale

The paper's central claim is that pLeiden preserves all six sequential Leiden guarantees. This rests on Lemmas 1 and 2, which assert that pLeidenR and LeidenR generate the same set of clusterings. The argument is synthetic, not circular: pLeidenR is defined as a spanning-tree-based procedure, and Lemma 1 explicitly simulates a LeidenR execution, while Lemma 2 attempts to construct the H and f inputs for pLeidenR from a given LeidenR execution. Neither lemma assumes the conclusion; the equivalence is the thing proved. The proof of Lemma 2 may contain logical gaps (e.g., the self-loop case sets H[v] = (u,u), but w'(u,u) < 0 makes that edge unavailable in pLeidenR), but a proof error is not an instance of circularity. The reported speedups are empirical comparisons against external baselines; the tunable parameters (phi and pass limit) are implementation choices with ablation studies, not parameters fitted to force the reported speedups. Self-citations to the authors' Jet paper and dataset are used as tooling and benchmarks, not as the load-bearing premise of the guarantee theorem. Therefore, the derivation chain does not reduce to its own inputs, and no significant circularity is present.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim of guarantee preservation relies on transferring sequential Leiden's proofs to the parallel setting, which requires the local-move and refinement traits to hold exactly. The parameters listed are implementation choices tuned on the test set, not fitted to produce the reported speedups.

free parameters (4)
  • local move pass limit = 6
    Set by the authors and used in all experiments; not derived from theory.
  • temperature parameter phi = 0.75 for passes 1-4, 0.25 for passes 5-6
    Controls the simulated annealing acceptance threshold; tuned for quality and speed, with ablation study.
  • epsilon tie-breaking threshold = 0.1
    Tolerance for treating objective deltas as equal in the afterburner filter ordering.
  • move fraction threshold = 0.05 (5%)
    Switches between updating and recomputing hash tables; chosen empirically.
assumptions (4)
  • domain assumption Original Leiden guarantees hold if the local move heuristic returns an improved clustering unless node optimal, and the refinement generates exactly the LeidenR clusterings.
    Invoked in Theorem 1 to transfer guarantees from sequential Leiden to pLeiden. This is the load-bearing bridge.
  • standard math The LambdaCC signed graph formulation preserves objective deltas for vertex moves, so w'(v, C) as defined tracks modularity/CPM.
    Used throughout Section II and in algorithm correctness checks; follows from Veldt et al.
  • domain assumption Random ordering functions on trees give expected logarithmic depth and allow edge subsampling to eliminate cycles.
    Used in pLeidenR construction (Section V-B). Not formally proven in this paper, only asserted.
  • domain assumption The synchronous bulk-synchronous parallel model on GPUs is a faithful basis for the local move heuristic; the afterburner filter's tie-breaking emulates an ordered traversal.
    Core to pLouvain's correctness and quality claims (Section IV-A).

how reviews work

0 comments
Cite this review

Pith. "Pith review of GPU-Accelerated Multilevel Graph Clustering: A Parallel Perspective on Louvain and Leiden." pith.science (2026). https://pith.science/paper/OXDW4ASQ

@misc{pith2026260801503,
  author       = {Pith},
  title        = {Pith review of: GPU-Accelerated Multilevel Graph Clustering: A Parallel Perspective on Louvain and Leiden},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OXDW4ASQ}},
  note         = {Machine review of arXiv:2608.01503}
}
read the original abstract

The sequential Louvain and Leiden algorithms are widely used techniques for modularity-optimizing clustering (or community detection) in large graphs. We present pLouvain and pLeiden, two new GPU parallelizations. pLouvain is based on the Louvain+ extension. pLeiden is the first parallel implementation to provably preserve all quality guarantees of sequential Leiden. We achieve this through a novel spanning-tree-based refinement approach. Both pLouvain and pLeiden use a lightweight symmetry-breaking technique that emulates an ordered traversal of vertices. For pLouvain, we develop an alternative iteration strategy to rectify the weak internal cluster connectivity observed in Louvain/Louvain+. Further, both pLouvain and pLeiden optimize the LambdaCC objective function, a generalization of modularity and the related Constant Potts model. On a collection of 57 graphs from 10 families, our results show that pLouvain and pLeiden achieve geometric mean speedups of 3.1x and 8.8x, respectively, over the current fastest open-source parallelizations of Louvain and Leiden. For the clusterings generated, pLouvain yields the highest modularity scores on nearly all tested graphs. The subroutines within these two multilevel approaches could aid in the parallelization of other Louvain-based techniques.

Figures

Figures reproduced from arXiv: 2608.01503 by the authors.

Figure 3
Figure 3. Running time (left) and modularity (right) comparisons with 10 [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 1
Figure 1. Running time comparison of clustering algorithms. [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Modularity comparison of clustering algorithms. [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    Community detection in graphs,

    S. Fortunato, “Community detection in graphs,”Physics reports, vol. 486, no. 3-5, pp. 75–174, 2010

    work page 2010

  2. [2]

    Fast unfolding of communities in large networks,

    V . D. Blondel, J.-L. Guillaume, R. Lambiotte, and E. Lefebvre, “Fast unfolding of communities in large networks,”J. Stat. Mech., vol. 2008, p. P10008, Oct. 2008

    work page 2008

  3. [3]

    From Louvain to Leiden: guaranteeing well-connected communities,

    V . A. Traag, L. Waltman, and N. J. van Eck, “From Louvain to Leiden: guaranteeing well-connected communities,”Sci. Rep., vol. 9, p. 5233, Mar. 2019

    work page 2019

  4. [4]

    Coarsening, sampling, and smoothing: elements of the multilevel method,

    S.-H. Teng, “Coarsening, sampling, and smoothing: elements of the multilevel method,” inAlgorithms for Parallel Processing(M. T. Heath, A. Ranade, and R. S. Schreiber, eds.), pp. 247–276, Springer, 1999

    work page 1999

  5. [5]

    High-performance imple- mentation of Louvain algorithm with representational optimizations,

    S. Sahu, K. Kothapalli, and D. S. Banerjee, “High-performance imple- mentation of Louvain algorithm with representational optimizations,” in Complex Networks & Their Applications XIII(H. Cherifi, M. Donduran, L. M. Rocha, C. Cherifi, and O. Varol, eds.), (Cham), pp. 127–139, Springer Nature Switzerland, 2025

    work page 2025

  6. [6]

    CPU vs. GPU for Community Detection: Performance Insights from GVE-Louvain and $\nu$-Louvain

    S. Sahu, “CPU vs. GPU for community detection: Performance insights from GVE-Louvain andν-Louvain,” 2025. arXiv:2501.19004

    work page Pith review arXiv 2025

  7. [7]

    Swift unfolding of communities: GPU-accelerated Louvain algorithm,

    Z. Wang, X. Lin, X. Li, P. Wang, Z. Meng, H. Liu, C. Tian, and S. Zhong, “Swift unfolding of communities: GPU-accelerated Louvain algorithm,” inProceedings of the 30th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming, PPoPP ’25, (New York, NY , USA), p. 441–454, Association for Computing Machinery, 2025

    work page 2025

  8. [8]

    Improving the Louvain algorithm for commu- nity detection with modularity maximization,

    O. Gach and J.-K. Hao, “Improving the Louvain algorithm for commu- nity detection with modularity maximization,” inArtificial Evolution (P. Legrand, M.-M. Corsini, J.-K. Hao, N. Monmarch ´e, E. Lutton, and M. Schoenauer, eds.), (Cham), pp. 145–156, Springer International Publishing, 2014

    work page 2014

Show all 41 references
  1. [9]

    Scalable community detection via parallel correlation clustering,

    J. Shi, L. Dhulipala, D. Eisenstat, J. Ł ˘acki, and V . Mirrokni, “Scalable community detection via parallel correlation clustering,”Proceedings of the VLDB Endowment, vol. 14, no. 11, pp. 2305–2313, 2021

    2021

  2. [10]

    Fast Leiden algorithm for community detection in shared memory setting,

    S. Sahu, K. Kothapalli, and D. S. Banerjee, “Fast Leiden algorithm for community detection in shared memory setting,” inProceedings of the 53rd International Conference on Parallel Processing, ICPP ’24, (New York, NY , USA), p. 11–20, Association for Computing Machinery, 2024

    2024

  3. [11]

    Jet: Multilevel graph partitioning on graphics processing units,

    M. S. Gilbert, K. Madduri, E. G. Boman, and S. Rajamanickam, “Jet: Multilevel graph partitioning on graphics processing units,”SIAM Journal on Scientific Computing, vol. 46, no. 5, pp. B700–B724, 2024

    2024

  4. [12]

    A correlation clustering framework for community detection,

    N. Veldt, D. F. Gleich, and A. Wirth, “A correlation clustering framework for community detection,” inProceedings of the 2018 World Wide Web Conference, pp. 439–448, 2018

    2018

  5. [13]

    Finding and evaluating community structure in networks,

    M. E. Newman and M. Girvan, “Finding and evaluating community structure in networks,”Physical review E, vol. 69, no. 2, p. 026113, 2004

    2004

  6. [14]

    Narrow scope for resolution-limit-free community detection,

    V . A. Traag, P. Van Dooren, and Y . Nesterov, “Narrow scope for resolution-limit-free community detection,”Phys. Rev. E, vol. 84, p. 016114, Jul 2011

    2011

  7. [15]

    Layered label propaga- tion: A multiresolution coordinate-free ordering for compressing social networks,

    P. Boldi, M. Rosa, M. Santini, and S. Vigna, “Layered label propaga- tion: A multiresolution coordinate-free ordering for compressing social networks,” inProc. 20th Int’l. Conf. on World Wide Web (WWW), 2011

    2011

  8. [16]

    Near linear time algorithm to detect community structures in large-scale networks,

    U. N. Raghavan, R. Albert, and S. Kumara, “Near linear time algorithm to detect community structures in large-scale networks,”Physical Review E, vol. 76, no. 3, p. 036106, 2007

    2007

  9. [17]

    On modularity clustering,

    U. Brandes, D. Delling, M. Gaertler, R. Gorke, M. Hoefer, Z. Nikoloski, and D. Wagner, “On modularity clustering,”IEEE Transactions on Knowledge and Data Engineering, vol. 20, no. 2, pp. 172–188, 2008

    2008

  10. [18]

    Memetic Graph Clustering,

    S. Biedermann, M. Henzinger, C. Schulz, and B. Schuster, “Memetic Graph Clustering,” inProceedings of the 17th International Symposium on Experimental Algorithms (SEA’18), LIPIcs, Dagstuhl, 2018. Techni- cal Report, arXiv:1802.07034

    2018 arXiv

  11. [19]

    Multilevel local search algorithms for modu- larity clustering,

    R. Rotta and A. Noack, “Multilevel local search algorithms for modu- larity clustering,”ACM J. Exp. Algorithmics, vol. 16, July 2011

    2011

  12. [20]

    Multilevel refinement for combinatorial optimisation prob- lems,

    C. Walshaw, “Multilevel refinement for combinatorial optimisation prob- lems,”Ann. Oper. Res., vol. 131, pp. 325–372, Oct. 2004

    2004

  13. [21]

    Isolate sets partition benefits community detection of parallel Louvain method,

    H. Qie, S. Li, Y . Dou, J. Xu, Y . Xiong, and Z. Gao, “Isolate sets partition benefits community detection of parallel Louvain method,”Sci. Rep., vol. 12, p. 8248, May 2022

    2022

  14. [22]

    Distributed Louvain algorithm for graph community detection,

    S. Ghosh, M. Halappanavar, A. Tumeo, A. Kalyanaraman, H. Lu, D. Chavarri `a-Miranda, A. Khan, and A. Gebremedhin, “Distributed Louvain algorithm for graph community detection,” in2018 IEEE International Parallel and Distributed Processing Symposium (IPDPS), pp. 885–895, 2018

    2018

  15. [23]

    Louvain community detection with parallel heuristics on GPUs,

    R. Forster, “Louvain community detection with parallel heuristics on GPUs,” in2016 IEEE 20th Jubilee International Conference on Intelli- gent Engineering Systems (INES), pp. 227–232, 2016

    2016

  16. [24]

    Batched graph community detection on GPUs,

    H.-Y . Chou and S. Ghosh, “Batched graph community detection on GPUs,” inProceedings of the International Conference on Parallel Architectures and Compilation Techniques, PACT ’22, (New York, NY , USA), p. 172–184, Association for Computing Machinery, 2023

    2023

  17. [25]

    Scalable static and dynamic community detection using grappolo,

    M. Halappanavar, H. Lu, A. Kalyanaraman, and A. Tumeo, “Scalable static and dynamic community detection using grappolo,” in2017 IEEE High Performance Extreme Computing Conference (HPEC), pp. 1–6, 2017

    2017

  18. [26]

    Community detection on the GPU,

    M. Naim, F. Manne, M. Halappanavar, and A. Tumeo, “Community detection on the GPU,” in2017 IEEE International Parallel and Dis- tributed Processing Symposium (IPDPS), pp. 625–634, 2017

    2017

  19. [27]

    Leiden-based parallel community detection,

    F. Nguyen, “Leiden-based parallel community detection,” Bachelor’s thesis, Karlsruhe Institute of Technology, 2021

    2021

  20. [28]

    Brief announcement: Distributed uncon- strained local search for multilevel graph partitioning,

    P. Sanders and D. Seemaier, “Brief announcement: Distributed uncon- strained local search for multilevel graph partitioning,” inProceedings of the 36th ACM Symposium on Parallelism in Algorithms and Archi- tectures, SPAA ’24, (New York, NY , USA), p. 443–445, Association for C...

    2024

  21. [29]

    Deterministic parallel high- quality hypergraph partitioning,

    R. Krause, L. Gottesb ¨uren, and N. Maas, “Deterministic parallel high- quality hypergraph partitioning,” in2025 Proceedings of the Conference on Applied and Computational Discrete Algorithms (ACDA), pp. 222– 236

  22. [30]

    Optimization techniques for GPU programming,

    P. Hijma, S. Heldens, A. Sclocco, B. Van Werkhoven, and H. E. Bal, “Optimization techniques for GPU programming,”ACM Computing Surveys, vol. 55, no. 11, pp. 1–81, 2023

    2023

  23. [31]

    Tera-scale multilevel graph partitioning,

    D. Salwasser, D. Seemaier, L. Gottesb ¨uren, and P. Sanders, “Tera-scale multilevel graph partitioning,” in2025 IEEE International Parallel and Distributed Processing Symposium (IPDPS), pp. 285–296, 2025

    2025

  24. [32]

    Scaling and validating louvain in cugraph against massive graphs,

    A. Zaheri and C. Hastings, “Scaling and validating louvain in cugraph against massive graphs,” 2022

    2022

  25. [33]

    How to accelerate community detection in python using gpu-powered leiden,

    R. Ratzel, “How to accelerate community detection in python using gpu-powered leiden,” 2025

    2025

  26. [34]

    Graph dataset for ‘Jet: Multilevel graph partitioning on graphics processing units’,

    M. Gilbert, K. Madduri, E. G. Boman, and S. Rajamanickam, “Graph dataset for ‘Jet: Multilevel graph partitioning on graphics processing units’,” 2024. doi:10.26207/pffm-mc36

    2024 doi

  27. [35]

    Uncovering and verifying optimal community structure in complex networks: A maxsat approach,

    C. Ans ´otegui, V . P. Ramaswamy, S. Szeider, and H. Xia, “Uncovering and verifying optimal community structure in complex networks: A maxsat approach,” inInternational Conference on Computational Sci- ence, pp. 35–49, Springer, 2025. Appendix: Artifact Description IX. OVERVIE...

    2025

  28. [36]

    v-Louvain (https://github.com/puzzlef/louvain- communities-cuda)

  29. [37]

    GVE Louvain (https://github.com/puzzlef/louvain- communities-openmp)

  30. [38]

    GVE Leiden (https://github.com/puzzlef/leiden- communities-openmp)

  31. [39]

    Networkit Louvain v11.1.post1 and Leiden v11.2 (https://networkit.github.io/)

  32. [40]

    cuGraph Louvain and Leiden (https://github.com/rapidsai/cugraph) v26.02

  33. [41]

    It contains a superset of the graphs listed in Table 1, represented in the Metis file format

    VieClus v1.1 (https://github.com/VieClus/VieClus) Datasets / Inputs:The test graph dataset is available at https://scholarsphere.psu.edu/resources/fd9ba209-a0cd-4f33- 994b-c22ae3bcb243/downloads/35163?download=true. It contains a superset of the graphs listed in Table 1, repre...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.