Khanh Nguyen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Khanh Nguyen has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Aidan D. Meade
Affiliations Radiation and Environmental Science Centre, Physical to Life Sciences Research Hub, Technological University Dublin, Dublin, Ireland, School of Physics, Clinical & Optometric Sciences, Technological University Dublin, Dublin, Ireland
16 shared publications - Barbara M. Ryan
Department of Gastroenterology, Tallaght Hospital, Dublin 24, Ireland
16 shared publications - Croí E. Buckley
Department of Surgery, School of Medicine, Trinity Translational Medicine Institute, Trinity College Dublin, Dublin, Ireland; Trinity St. James's Cancer Institute, St. James's Hospital, Trinity College Dublin, Dublin, Ireland
16 shared publications - David Mockler
Dublin Trinity College Dublin School of Medicine
16 shared publications - Eimear Mylod
Cancer Immunology and Immunotherapy Group, Department of Surgery, Trinity Translational Medicine Institute, Trinity St James’s Cancer Institute, St. James’s Hospital, Trinity College Dublin, Dublin, Ireland
16 shared publications - Fiona O&Apos; Connell
Department of Surgery, Trinity St. James’s Cancer Institute and Trinity Translational Medicine Institute, St. James’s Hospital and Trinity College Dublin, D08 W9RT Dublin, Ireland
16 shared publications
Latest Publications
- Open Access
- 1 October 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Programming Languages
- Vol. 10 (OOPSLA2) , 1766-1793
Bonsai: Efficient and Optimal Automatic Tensor Rematerialization for Memory-Constrained DNN Training
GPU memory is increasingly the primary bottleneck in scaling deep neural network (DNN) training, where the activation tensors footprint of a model may exceed the memory capacity. Tensor recomputation is a powerful technique that trades additional computation for reduced peak memory usage. However, existing approaches face a fundamental tension between performance optimality and computational scalability. On the one hand, solvers leverage Integer Linear Programming (ILP) to provide mathematically optimal solutions but suffer from the combinatorial explosion of the search space and thus become intractable for modern DNN models. On the other hand, heuristics-based approaches achieve scalability but sacrifice optimality altogether, resulting in suboptimal execution schedules. The root cause of these inefficiencies in the state of the art is the mismatch in abstraction. This paper introduces Bonsai, a framework that tackles this scalability-granularity tension. At the heart of Bonsai is a novel abstraction of operator segmentation that breaks the computation graph into flexible, variable-sized units to enable a lightweight yet effective segment-based ILP formulation. By having segments, Bonsai collapses the search space and prunes redundant solutions that stall existing solvers. This abstraction enables Bonsai to maintain a holistic view of the entire model, ensuring that no optimization opportunity is lost while reducing the number of decision variables by orders of magnitude. The evaluation across a diverse set of DNN architectures and models demonstrates that Bonsai scales to real-world models, is up to 10.13× lower solver cost than state-of-the-art ILP solvers, and delivers up to 11
- 1 October 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Programming Languages
- Vol. 10 (OOPSLA2) , 1766-1793
Open AccessBonsai: Efficient and Optimal Automatic Tensor Rematerialization for Memory-Constrained DNN Training
GPU memory is increasingly the primary bottleneck in scaling deep neural network (DNN) training, where the activation tensors footprint of a model may exceed the memory capacity. Tensor recomputation is a powerful technique that trades additional computation for reduced peak memory usage. However, existing approaches face a fundamental tension between performance optimality and computational scalability. On the one hand, solvers leverage Integer Linear Programming (ILP) to provide mathematically optimal solutions but suffer from the combinatorial explosion of the search space and thus become intractable for modern DNN models. On the other hand, heuristics-based approaches achieve scalability but sacrifice optimality altogether, resulting in suboptimal execution schedules. The root cause of these inefficiencies in the state of the art is the mismatch in abstraction. This paper introduces Bonsai, a framework that tackles this scalability-granularity tension. At the heart of Bonsai is a novel abstraction of operator segmentation that breaks the computation graph into flexible, variable-sized units to enable a lightweight yet effective segment-based ILP formulation. By having segments, Bonsai collapses the search space and prunes redundant solutions that stall existing solvers. This abstraction enables Bonsai to maintain a holistic view of the entire model, ensuring that no optimization opportunity is lost while reducing the number of decision variables by orders of magnitude. The evaluation across a diverse set of DNN architectures and models demonstrates that Bonsai scales to real-world models, is up to 10.13× lower solver cost than state-of-the-art ILP solvers, and delivers up to 11
- Open Access
- 11 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 91-97
Meshilon: A Merging-based Java Garbage Collector
- 11 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 91-97
Open AccessMeshilon: A Merging-based Java Garbage Collector
- Open Access
- 25 February 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-12
Enhancing Program Analysis with Deterministic Distinguishable Calling Context
Calling context is crucial for improving the precision of program analyses in various use cases (clients), such as profiling, debugging, optimization, and security checking. Often the calling context is encoded using a numerical value. We have observed that many clients benefit not only from a deterministic but also globally distinguishable value across runs to simplify bookkeeping and guarantee complete uniqueness. However, existing work only guarantees determinism, not global distinguishability. Clients need to develop auxiliary helpers, which incurs considerable overhead to distinguish encoded values among all calling contexts. In this paper, we propose Deterministic Distinguishable Calling Context Encoding () that can enable both properties of calling context encoding natively. The key idea of is leveraging the static call graph and encoding each calling context as the running call path count. Thereby, a mapping is established statically and can be readily used by the clients. Our experiments with two client tools show that has a comparable overhead compared to two state-of-the-art encoding schemes, PCCE and PCC, and further avoids the expensive overheads of collision detection, up to 2.1× and 50%, for Splash-3 and SPEC CPU 2017, respectively.
- 25 February 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-12
Open AccessEnhancing Program Analysis with Deterministic Distinguishable Calling Context
Calling context is crucial for improving the precision of program analyses in various use cases (clients), such as profiling, debugging, optimization, and security checking. Often the calling context is encoded using a numerical value. We have observed that many clients benefit not only from a deterministic but also globally distinguishable value across runs to simplify bookkeeping and guarantee complete uniqueness. However, existing work only guarantees determinism, not global distinguishability. Clients need to develop auxiliary helpers, which incurs considerable overhead to distinguish encoded values among all calling contexts. In this paper, we propose Deterministic Distinguishable Calling Context Encoding () that can enable both properties of calling context encoding natively. The key idea of is leveraging the static call graph and encoding each calling context as the running call path count. Thereby, a mapping is established statically and can be readily used by the clients. Our experiments with two client tools show that has a comparable overhead compared to two state-of-the-art encoding schemes, PCCE and PCC, and further avoids the expensive overheads of collision detection, up to 2.1× and 50%, for Splash-3 and SPEC CPU 2017, respectively.
- Open Access
- 4 September 2024
- Published by Association for Computing Machinery (ACM)
- p. 15-22
Polar: A Managed Runtime with Hotness-Segregated Heap for Far Memory
Thanks to recent advances in high-bandwidth, low-latency interconnects, running a data-intensive application with a remote memory pool is now a feasibility. When developing a data-intensive application, a managed language such as Java is often the developer's choice due to convenience of the runtime such as automatic memory management. However, the memory management cost increases significantly in far memory due to remote memory accesses. Our insight is that data hotness (i.e., access frequency of objects) is the key to reducing the memory management cost and improving efficiency in far memory. In this paper, we present an ongoing work designing Polar, an enhanced runtime system that is hotness-aware, and optimized for far memory. In Polar, the garbage collector is augmented to identify cold (infrequently accessed) objects and relocate them to remote memory pools. By placing objects at memory locations based on their access frequency, Polar minimizes the number of remote accesses, ensures low access latency for the application, and thus improves overall performance.
- 4 September 2024
- Published by Association for Computing Machinery (ACM)
- p. 15-22
Open AccessPolar: A Managed Runtime with Hotness-Segregated Heap for Far Memory
Thanks to recent advances in high-bandwidth, low-latency interconnects, running a data-intensive application with a remote memory pool is now a feasibility. When developing a data-intensive application, a managed language such as Java is often the developer's choice due to convenience of the runtime such as automatic memory management. However, the memory management cost increases significantly in far memory due to remote memory accesses. Our insight is that data hotness (i.e., access frequency of objects) is the key to reducing the memory management cost and improving efficiency in far memory. In this paper, we present an ongoing work designing Polar, an enhanced runtime system that is hotness-aware, and optimized for far memory. In Polar, the garbage collector is augmented to identify cold (infrequently accessed) objects and relocate them to remote memory pools. By placing objects at memory locations based on their access frequency, Polar minimizes the number of remote accesses, ensures low access latency for the application, and thus improves overall performance.
- Open Access
- 4 April 2023
- Published by American Association for Cancer Research (AACR)
Supplementary Material S3 from Diagnostic Accuracy of Blood-based Biomarkers for Pancreatic Cancer: A Systematic Review and Meta-analysis
Individualised search strategy
- 4 April 2023
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Material S3 from Diagnostic Accuracy of Blood-based Biomarkers for Pancreatic Cancer: A Systematic Review and Meta-analysis
Individualised search strategy