Yunfei Pang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yunfei Pang has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Guangming Tan
Chinese Academy of Sciences
3 shared publications - Junmin Xiao
Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
3 shared publications - Mingyi Li
Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
3 shared publications - Chaoyang Shui
Institute of Computing Technology, Chinese Academy of Sciences, China and University of Chinese Academy of Sciences, China
2 shared publications - Hui Ma
Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
2 shared publications - Zehua Wang
Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
2 shared publications
Latest Publications
- Open Access
- 22 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 1949-1965
T-Control: An Efficient Dynamic Tensor Rematerialization System for DNN Training
With the continuous growth of model and batch sizes, DNN model training increasingly suffers from excessive memory consumption. Tensor rematerialization has emerged as an effective technique to enable training under limited memory constraints. However, dynamic rematerialization methods often underperform static approaches, primarily due to their greedy tensor eviction schedules and runtime overhead. In this work, we develop T-Control, a dynamic tensor rematerialization system for DNN model training. T-Control integrates the topology of the traced tensor dependency graph with real-time memory usage to make informed, adaptive tensor retention decisions, preserving critical tensors and reducing eviction-induced recomputation. Furthermore, by extending PyTorch's native memory manager with our fine-grained memory management strategy, T-control effectively improves memory utilization, thereby reducing unnecessary tensor eviction. Experimental results demonstrate that T-Control boosts throughput by up to 1.58× and 1.91× over state-of-the-art static and dynamic tensor rematerialization systems.
- 22 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 1949-1965
Open AccessT-Control: An Efficient Dynamic Tensor Rematerialization System for DNN Training
With the continuous growth of model and batch sizes, DNN model training increasingly suffers from excessive memory consumption. Tensor rematerialization has emerged as an effective technique to enable training under limited memory constraints. However, dynamic rematerialization methods often underperform static approaches, primarily due to their greedy tensor eviction schedules and runtime overhead. In this work, we develop T-Control, a dynamic tensor rematerialization system for DNN model training. T-Control integrates the topology of the traced tensor dependency graph with real-time memory usage to make informed, adaptive tensor retention decisions, preserving critical tensors and reducing eviction-induced recomputation. Furthermore, by extending PyTorch's native memory manager with our fine-grained memory management strategy, T-control effectively improves memory utilization, thereby reducing unnecessary tensor eviction. Experimental results demonstrate that T-Control boosts throughput by up to 1.58× and 1.91× over state-of-the-art static and dynamic tensor rematerialization systems.
- Open Access
- 30 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 460-472
A Coordinated Strategy for GNN Combining Computational Graph and Operator Optimizations
Graph Neural Networks (GNNs) have garnered significant interest across various domains due to their efficacy in learning from graph-structured data. In pursuit of heightened performance, numerous GNN frameworks have emerged recently. However, recent work tends to study performance optimization at the computational graph level and operator level separately, and the existing optimization techniques rely on pattern matching and manual intervention, driven by human expertise. Consequently, their performances remain sub-optimal and sensitive to input graphs and GNN models. In this work, we develop an efficient coordinated strategy named AlphaGNN, which achieves an effective combination of computational graph optimization and operator optimization. To render this coordinated optimization impactful, a rule-based computational graph optimization and a performance-driven operator optimization are proposed. The experimental results confirm that AlphaGNN achieves up to 12.39 × (2.94 × on average) performance improvement over the state-of-the-art methods on diverse GNN models.
- 30 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 460-472
Open AccessA Coordinated Strategy for GNN Combining Computational Graph and Operator Optimizations
Graph Neural Networks (GNNs) have garnered significant interest across various domains due to their efficacy in learning from graph-structured data. In pursuit of heightened performance, numerous GNN frameworks have emerged recently. However, recent work tends to study performance optimization at the computational graph level and operator level separately, and the existing optimization techniques rely on pattern matching and manual intervention, driven by human expertise. Consequently, their performances remain sub-optimal and sensitive to input graphs and GNN models. In this work, we develop an efficient coordinated strategy named AlphaGNN, which achieves an effective combination of computational graph optimization and operator optimization. To render this coordinated optimization impactful, a rule-based computational graph optimization and a performance-driven operator optimization are proposed. The experimental results confirm that AlphaGNN achieves up to 12.39 × (2.94 × on average) performance improvement over the state-of-the-art methods on diverse GNN models.
- Open Access
- 11 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-15
Adaptive Workload-Balanced Scheduling Strategy for Global Ocean Data Assimilation on Massive GPUs
Global ocean data assimilation is a crucial technique to estimate the actual oceanic state by combining numerical model outcomes and observation data, which is widely used in climate research. Due to the imbalanced distribution of observation data in global ocean, the parallel efficiency of recent methods suffers from workload imbalance. When massive GPUs are applied for global ocean data assimilation, the workload imbalance becomes more severe, resulting in poor scalability. In this work, we propose a novel adaptive workload-balance scheduling strategy, Bassimilation, which successfully estimates the total workload prior to execution and ensures a balanced workload assignment. Further, we design a parallel dynamic programming approach to accelerate the schedule decision, and develop a factored dataflow to exploit the parallel potential of GPUs. Evaluation demonstrates that our algorithm outperforms the state-of-the-art method by up to 9.1× speedup. This work is the first to scale global ocean data assimilation to 4, 000 GPUs.
- 11 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-15
Open AccessAdaptive Workload-Balanced Scheduling Strategy for Global Ocean Data Assimilation on Massive GPUs
Global ocean data assimilation is a crucial technique to estimate the actual oceanic state by combining numerical model outcomes and observation data, which is widely used in climate research. Due to the imbalanced distribution of observation data in global ocean, the parallel efficiency of recent methods suffers from workload imbalance. When massive GPUs are applied for global ocean data assimilation, the workload imbalance becomes more severe, resulting in poor scalability. In this work, we propose a novel adaptive workload-balance scheduling strategy, Bassimilation, which successfully estimates the total workload prior to execution and ensures a balanced workload assignment. Further, we design a parallel dynamic programming approach to accelerate the schedule decision, and develop a factored dataflow to exploit the parallel potential of GPUs. Evaluation demonstrates that our algorithm outperforms the state-of-the-art method by up to 9.1× speedup. This work is the first to scale global ocean data assimilation to 4, 000 GPUs.