David F. Gleich
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
David F. Gleich has a total of 133 co-authors. Here are the top co-authors by number of shared publications:
- Kyle Kloster
Computer Science Mathematics Computer Science} Department Purdue University
4 shared publications - Michael W. Mahoney
Department of Statistics ICSI LBNL University of California at Berkeley
4 shared publications - Ananth Grama
Purdue University, West Lafayette, IN, USA
3 shared publications - Andrew A. Chien
University of Chicago, Chicago, USA
3 shared publications - Andronicus Rajasukumar
University of Chicago, Chicago, USA
3 shared publications - Austin R. Benson
Cornell University
3 shared publications
Latest Publications
- Open Access
- 5 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 603-617
UpDown: Efficient Manycore based on Many Threading and Scalable Memory Parallelism
- 5 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 603-617
Open AccessUpDown: Efficient Manycore based on Many Threading and Scalable Memory Parallelism
- Open Access
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 37 (7) , 1553-1570
UpDown: A Supercomputer Co-Designed for Scalable Graph Processing
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 37 (7) , 1553-1570
Open AccessUpDown: A Supercomputer Co-Designed for Scalable Graph Processing
- Open Access
- 19 March 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (12)
Powers of magnetic graph matrix: Fourier spectrum, walk compression, and applications
Magnetic graphs, originally developed to model quantum systems under magnetic fields, have recently emerged as a powerful framework for analyzing complex directed networks. Existing research has primarily used the spectral properties of the magnetic graph matrix to study global and stationary network features. However, their capacity to model local, nonequilibrium behaviors, often described by matrix powers, remains largely unexplored. We present a combinatorial interpretation of the magnetic graph matrix powers through directed walk profiles—counts of graph walks indexed by the number of edge reversals. Crucially, we establish that walk profiles correspond to a Fourier transform of magnetic matrix powers. The connection allows exact reconstruction of walk profiles from magnetic matrix powers at multiple discrete potentials, and more importantly, an even smaller number of potentials often suffices for accurate approximate reconstruction in real networks. This shows the empirical compressibility of the information captured by the magnetic matrix. This fresh perspective suggests further applications; for example, we illustrate how powers of the magnetic matrix can identify frustrated directed cycles (e.g., feedforward loops) and can be effectively employed for link prediction by encoding local structural details in directed graphs.
- 19 March 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (12)
Open AccessPowers of magnetic graph matrix: Fourier spectrum, walk compression, and applications
Magnetic graphs, originally developed to model quantum systems under magnetic fields, have recently emerged as a powerful framework for analyzing complex directed networks. Existing research has primarily used the spectral properties of the magnetic graph matrix to study global and stationary network features. However, their capacity to model local, nonequilibrium behaviors, often described by matrix powers, remains largely unexplored. We present a combinatorial interpretation of the magnetic graph matrix powers through directed walk profiles—counts of graph walks indexed by the number of edge reversals. Crucially, we establish that walk profiles correspond to a Fourier transform of magnetic matrix powers. The connection allows exact reconstruction of walk profiles from magnetic matrix powers at multiple discrete potentials, and more importantly, an even smaller number of potentials often suffices for accurate approximate reconstruction in real networks. This shows the empirical compressibility of the information captured by the magnetic matrix. This fresh perspective suggests further applications; for example, we illustrate how powers of the magnetic matrix can identify frustrated directed cycles (e.g., feedforward loops) and can be effectively employed for link prediction by encoding local structural details in directed graphs.
- Open Access
Decoupling Variance and Scale-Invariant Updates in Adaptive Gradient Descent for Unified Vector and Matrix Optimization
Adaptive methods like Adam have become the $\textit{de facto}$ standard for large-scale vector and Euclidean optimization due to their coordinate-wise adaptation with a second-order nature. More recently, matrix-based spectral optimizers like Muon (Jordan et al., 2024b) show the power of treating weight matrices as matrices rather than long vectors. Linking these is hard because many natural generalizations are not feasible to implement, and we also cannot simply move the Adam adaptation to the matrix spectrum. To address this, we reformulate the AdaGrad update and decompose it into a variance adaptation term and a scale-invariant term. This decoupling produces $\textbf{DeVA}$ ($\textbf{De}$coupled $\textbf{V}$ariance $\textbf{A}$daptation), a framework that bridges between vector-based variance adaptation and matrix spectral optimization, enabling a seamless transition from Adam to adaptive spectral descent. Extensive experiments across language modeling and image classification demonstrate that DeVA consistently outperforms state-of-the-art methods such as Muon and SOAP (Vyas et al., 2024), reducing token usage by around 6.6\%. Theoretically, we show that the variance adaptation term effectively improves the blockwise smoothness, facilitating faster convergence. Our implementation is available at https://github.com/Tsedao/Decoupled-Variance-AdaptationOpen AccessDecoupling Variance and Scale-Invariant Updates in Adaptive Gradient Descent for Unified Vector and Matrix Optimization
Adaptive methods like Adam have become the $\textit{de facto}$ standard for large-scale vector and Euclidean optimization due to their coordinate-wise adaptation with a second-order nature. More recently, matrix-based spectral optimizers like Muon (Jordan et al., 2024b) show the power of treating weight matrices as matrices rather than long vectors. Linking these is hard because many natural generalizations are not feasible to implement, and we also cannot simply move the Adam adaptation to the matrix spectrum. To address this, we reformulate the AdaGrad update and decompose it into a variance adaptation term and a scale-invariant term. This decoupling produces $\textbf{DeVA}$ ($\textbf{De}$coupled $\textbf{V}$ariance $\textbf{A}$daptation), a framework that bridges between vector-based variance adaptation and matrix spectral optimization, enabling a seamless transition from Adam to adaptive spectral descent. Extensive experiments across language modeling and image classification demonstrate that DeVA consistently outperforms state-of-the-art methods such as Muon and SOAP (Vyas et al., 2024), reducing token usage by around 6.6\%. Theoretically, we show that the variance adaptation term effectively improves the blockwise smoothness, facilitating faster convergence. Our implementation is available at https://github.com/Tsedao/Decoupled-Variance-Adaptation - Open Access
Fault Oblivious Eigenvalue Solver
Eigenvalue problems serve as fundamental substrates for applications in large-scale scientific simulations and machine learning, often requiring computation on massively parallel platforms. As these platforms scale to hundreds of thousands of cores, hardware failures become a significant challenge to reliability and efficiency. In this paper, we propose and analyze a novel fault-tolerant eigenvalue solver based on erasure-coded computations -- a technique that enhances resilience by augmenting the system with redundant data a priori. This transformation reformulates the original eigenvalue problem as a generalized eigenvalue problem, enabling fault-oblivious computation while preserving numerical stability and convergence properties. We formulate the augmentation scheme, establish the necessary conditions for the encoded blocks, and prove the relationship between the original and transformed problems. We implement an erasure-coded TraceMin eigensolver and demonstrate its effectiveness in extracting eigenvalues in the presence of faults. Our experimental results show that the proposed solver incurs minimal computational overhead, maintains robust convergence, and scales efficiently with the number of faults, making it a practical solution for resilient eigenvalue computations in large-scale systems.
Open AccessFault Oblivious Eigenvalue Solver
Eigenvalue problems serve as fundamental substrates for applications in large-scale scientific simulations and machine learning, often requiring computation on massively parallel platforms. As these platforms scale to hundreds of thousands of cores, hardware failures become a significant challenge to reliability and efficiency. In this paper, we propose and analyze a novel fault-tolerant eigenvalue solver based on erasure-coded computations -- a technique that enhances resilience by augmenting the system with redundant data a priori. This transformation reformulates the original eigenvalue problem as a generalized eigenvalue problem, enabling fault-oblivious computation while preserving numerical stability and convergence properties. We formulate the augmentation scheme, establish the necessary conditions for the encoded blocks, and prove the relationship between the original and transformed problems. We implement an erasure-coded TraceMin eigensolver and demonstrate its effectiveness in extracting eigenvalues in the presence of faults. Our experimental results show that the proposed solver incurs minimal computational overhead, maintains robust convergence, and scales efficiently with the number of faults, making it a practical solution for resilient eigenvalue computations in large-scale systems.
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1243-1262
KVMSR+UDWeave: Extreme-Scaling with Fine-grained Parallelism on the UpDown Graph Supercomputer
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1243-1262
Open AccessKVMSR+UDWeave: Extreme-Scaling with Fine-grained Parallelism on the UpDown Graph Supercomputer
- 1 November 2025
- Published by Springer Nature
- p. 78-95
Efficiently Exploiting Irregular Parallelism Using Keys at Scale
- 1 November 2025
- Published by Springer Nature
- p. 78-95
Efficiently Exploiting Irregular Parallelism Using Keys at Scale
- Open Access
How Fast Can Graph Computations Go on Fine-grained Parallel Architectures
Large-scale graph problems are of critical and growing importance and historically parallel architectures have provided little support. In the spirit of co-design, we explore the question, How fast can graph computing go on a fine-grained architecture? We explore the possibilities of an architecture optimized for fine-grained parallelism, natural programming, and the irregularity and skew found in real-world graphs. Using two graph benchmarks, PageRank (PR) and Breadth-First Search (BFS), we evaluate a Fine-Grained Graph architecture, UpDown, to explore what performance codesign can achieve. To demonstrate programmability, we wrote five variants of these algorithms. Simulations of up to 256 nodes (524,288 lanes) and projections to 16,384 nodes (33M lanes) show the UpDown system can achieve 637K GTEPS PR and 989K GTEPS BFS on RMAT, exceeding the best prior results by 5x and 100x respectively.
Open AccessHow Fast Can Graph Computations Go on Fine-grained Parallel Architectures
Large-scale graph problems are of critical and growing importance and historically parallel architectures have provided little support. In the spirit of co-design, we explore the question, How fast can graph computing go on a fine-grained architecture? We explore the possibilities of an architecture optimized for fine-grained parallelism, natural programming, and the irregularity and skew found in real-world graphs. Using two graph benchmarks, PageRank (PR) and Breadth-First Search (BFS), we evaluate a Fine-Grained Graph architecture, UpDown, to explore what performance codesign can achieve. To demonstrate programmability, we wrote five variants of these algorithms. Simulations of up to 256 nodes (524,288 lanes) and projections to 16,384 nodes (33M lanes) show the UpDown system can achieve 637K GTEPS PR and 989K GTEPS BFS on RMAT, exceeding the best prior results by 5x and 100x respectively.
- 29 March 2025
- Published by Springer Nature
- p. 43-55
A Spatial Hypergraph Model Where Epidemic Spread Demonstrates Clear Higher-Order Effects
- 29 March 2025
- Published by Springer Nature
- p. 43-55
A Spatial Hypergraph Model Where Epidemic Spread Demonstrates Clear Higher-Order Effects
- Open Access
- 11 March 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Framework for De novo Sequencing of Peptide Mixtures via Network Analysis and Two-Dimensional Tandem Mass Spectrometry
Two-dimensional tandem mass spectrometry (2D MS/MS) provides in-depth biopolymer structural information previously not directly accessible with traditional one-dimensional MS/MS workflows, and in significantly less time (<1 second per sample). In this study, we enhance analysis of 2D MS/MS data for greater applicability in omics workflows and address challenges in sequencing peptide of mixtures. We designed a graph-theory-based framework to efficiently manage, visualize, and maximize the structural information extractable from 2D MS/MS spectra. Graph analysis algorithms, including a PageRank-based method, are shown to deconvolve MS/MS signals and group together product ions from the same peptide precursor, enabling the reconstruction of peptide fragmentation trees. From this, MSn information can be extracted to improve sequencing accuracy relative to current MS/MS methods. We also introduce a computationally efficient de novo sequencing approach that leverages this structural information to reduce reliance on databases and sample separation, while also enabling the rapid sequencing of post-translationally modified peptides. Tests on simulated 2D MS/MS spectra, designed to mimic those from proteomic samples, achieved high precision in signal assignment. Proof-of-concept studies were conducted on real data from simple mixtures of short chain peptides, showing the potential applicability of combining network analysis with de novo sequencing to analyze unknown peptide mixtures. We anticipate that this technique will complement proteomics workflows and facilitate direct biopolymer structural analysis.
- 11 March 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessFramework for De novo Sequencing of Peptide Mixtures via Network Analysis and Two-Dimensional Tandem Mass Spectrometry
Two-dimensional tandem mass spectrometry (2D MS/MS) provides in-depth biopolymer structural information previously not directly accessible with traditional one-dimensional MS/MS workflows, and in significantly less time (<1 second per sample). In this study, we enhance analysis of 2D MS/MS data for greater applicability in omics workflows and address challenges in sequencing peptide of mixtures. We designed a graph-theory-based framework to efficiently manage, visualize, and maximize the structural information extractable from 2D MS/MS spectra. Graph analysis algorithms, including a PageRank-based method, are shown to deconvolve MS/MS signals and group together product ions from the same peptide precursor, enabling the reconstruction of peptide fragmentation trees. From this, MSn information can be extracted to improve sequencing accuracy relative to current MS/MS methods. We also introduce a computationally efficient de novo sequencing approach that leverages this structural information to reduce reliance on databases and sample separation, while also enabling the rapid sequencing of post-translationally modified peptides. Tests on simulated 2D MS/MS spectra, designed to mimic those from proteomic samples, achieved high precision in signal assignment. Proof-of-concept studies were conducted on real data from simple mixtures of short chain peptides, showing the potential applicability of combining network analysis with de novo sequencing to analyze unknown peptide mixtures. We anticipate that this technique will complement proteomics workflows and facilitate direct biopolymer structural analysis.