Junsheng Chang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Junsheng Chang has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Li Shen
National University of Defense Technology
4 shared publications - Shi Chen
National University of Defense Technology
2 shared publications - Yang Guo
National University of Defense Technology
2 shared publications Biwei Zhang
School of Computer, National University of Defense Technology, Key Laboratory of Advanced Microprocessors Chips and Systems, Changsha, China
1 shared publication- DongHuan Xie
College of Computer Science and Technology, National University of Defense Technology, Changsha, China
1 shared publication Fei Lei
School of Computer, National University of Defense Technology, Key Laboratory of Advanced Microprocessors Chips and Systems, Changsha, China
1 shared publication
Latest Publications
- Open Access
- 13 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 23 (4) , 1-24
PipeGNN: A Bandwidth-Efficient GNN Accelerator with Node-Level Pipelined Push Execution
Graph Neural Networks (GNNs) have become a fundamental tool for learning over graph-structured data. Under the message-passing framework, mainstream GNN models alternate between feature transformation and neighborhood aggregation. Fusing these two phases into a node-level pipelined push dataflow, in which each node’s transformed feature streams directly into aggregation, establishes a single-pass I/O target that reads each input tensor exactly once per layer. No existing hardware sustains this dataflow at full concurrency. Prior fusion accelerators fall short of this target in three ways: designs that gather the raw input features for aggregation or write transformed features off-chip incur redundant traffic, designs that stream transformed features on-chip still keep the two phases as separate matrix kernels without a unified execution granularity, and the designs that aggregate at message granularity serialize per-destination aggregation or allocate state that does not scale to large graphs. We propose PipeGNN, a bandwidth-efficient GNN accelerator whose streaming microarchitecture sustains this dataflow as a fully concurrent pipeline within a fixed on-chip buffer budget. PipeGNN decouples execution from topological boundaries by aligning execution granularity with both memory burst size and aggregation compute throughput, and resolves four coupled microarchitectural barriers through coordinated mechanisms: compute-aligned edge batching, contention-aware message reduction, adaptive cross-batch scheduling, and tiered state residency. Evaluated across diverse GNN workloads, PipeGNN achieves geometric mean speedups of 10.1× over AWB-GCN and 12.2× over FlowGNN, and 2.4× to 46.9× over a GPU baseline across graph scales, reduces off-chip traffic by 2.69× over AWB-GCN and 5.93× over FlowGNN, and improves energy efficiency by 56.81× over AWB-GCN and 1.53× over FlowGNN, with a 154× traffic reduction and a 116.81× energy reduction against HyGCN.
- 13 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 23 (4) , 1-24
Open AccessPipeGNN: A Bandwidth-Efficient GNN Accelerator with Node-Level Pipelined Push Execution
Graph Neural Networks (GNNs) have become a fundamental tool for learning over graph-structured data. Under the message-passing framework, mainstream GNN models alternate between feature transformation and neighborhood aggregation. Fusing these two phases into a node-level pipelined push dataflow, in which each node’s transformed feature streams directly into aggregation, establishes a single-pass I/O target that reads each input tensor exactly once per layer. No existing hardware sustains this dataflow at full concurrency. Prior fusion accelerators fall short of this target in three ways: designs that gather the raw input features for aggregation or write transformed features off-chip incur redundant traffic, designs that stream transformed features on-chip still keep the two phases as separate matrix kernels without a unified execution granularity, and the designs that aggregate at message granularity serialize per-destination aggregation or allocate state that does not scale to large graphs. We propose PipeGNN, a bandwidth-efficient GNN accelerator whose streaming microarchitecture sustains this dataflow as a fully concurrent pipeline within a fixed on-chip buffer budget. PipeGNN decouples execution from topological boundaries by aligning execution granularity with both memory burst size and aggregation compute throughput, and resolves four coupled microarchitectural barriers through coordinated mechanisms: compute-aligned edge batching, contention-aware message reduction, adaptive cross-batch scheduling, and tiered state residency. Evaluated across diverse GNN workloads, PipeGNN achieves geometric mean speedups of 10.1× over AWB-GCN and 12.2× over FlowGNN, and 2.4× to 46.9× over a GPU baseline across graph scales, reduces off-chip traffic by 2.69× over AWB-GCN and 5.93× over FlowGNN, and improves energy efficiency by 56.81× over AWB-GCN and 1.53× over FlowGNN, with a 154× traffic reduction and a 116.81× energy reduction against HyGCN.
- Open Access
- 24 September 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
DCSR-GCN: A High-Performance GCN Accelerator Based on Dynamic Compression and Sparsity Reordering
Graph Convolutional Networks (GCNs) are widely used in tasks involving irregular graph data, such as recommendation. The hybrid execution pattern of sparse aggregation and dense combination during inference limits the efficiency of general processors like CPU and GPU. Therefore, designing dedicated accelerators for GCN inference has become an important research direction. However, existing GCN accelerators still suffer from high storage overhead caused by sparse indices and frequent read-after-write (RAW) conflicts during the aggregation phase, which limit system throughput and energy efficiency. To address these challenges, we adopt the hardware-software co-design approach and present DCSR-GCN, a high-performance GCN accelerator based on dynamic compression and sparsity reordering. First, we design an Enhanced Compressed Sparse Row (ECSR) format to reduce adjacency index storage space, and directly embed node degree fields to eliminate redundant computations in aggregation. Second, we propose a two-phase reordering algorithm that combines conflict-aware row scheduling with reuse-aware column grouping, thereby mitigating RAW conflicts and improving memory access efficiency. Finally, to fully exploit the benefits of sparse optimizations, we design a high-throughput vector systolic array that accelerates feature transformation in combination. Evaluations show that DCSR-GCN achieves up to 26.03 ×, 4.19 ×, and 3.52 × speedup compared to HyGCN, GCNAX, and MEGA, respectively.
- 24 September 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
Open AccessDCSR-GCN: A High-Performance GCN Accelerator Based on Dynamic Compression and Sparsity Reordering
Graph Convolutional Networks (GCNs) are widely used in tasks involving irregular graph data, such as recommendation. The hybrid execution pattern of sparse aggregation and dense combination during inference limits the efficiency of general processors like CPU and GPU. Therefore, designing dedicated accelerators for GCN inference has become an important research direction. However, existing GCN accelerators still suffer from high storage overhead caused by sparse indices and frequent read-after-write (RAW) conflicts during the aggregation phase, which limit system throughput and energy efficiency. To address these challenges, we adopt the hardware-software co-design approach and present DCSR-GCN, a high-performance GCN accelerator based on dynamic compression and sparsity reordering. First, we design an Enhanced Compressed Sparse Row (ECSR) format to reduce adjacency index storage space, and directly embed node degree fields to eliminate redundant computations in aggregation. Second, we propose a two-phase reordering algorithm that combines conflict-aware row scheduling with reuse-aware column grouping, thereby mitigating RAW conflicts and improving memory access efficiency. Finally, to fully exploit the benefits of sparse optimizations, we design a high-throughput vector systolic array that accelerates feature transformation in combination. Evaluations show that DCSR-GCN achieves up to 26.03 ×, 4.19 ×, and 3.52 × speedup compared to HyGCN, GCNAX, and MEGA, respectively.
- Open Access
- 19 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 290-298
Row-wise Inter-Phase Pipelining for Hardware-Efficient GCN Acceleration
- 19 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 290-298
Open AccessRow-wise Inter-Phase Pipelining for Hardware-Efficient GCN Acceleration
- Open Access
- 28 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 98-106
Combination of Storage and Accumulation for Synchronous SpMV Acceleration on FPGAs with HBM
- 28 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 98-106
Open AccessCombination of Storage and Accumulation for Synchronous SpMV Acceleration on FPGAs with HBM
- 23 May 2024
- Published by Springer Nature in CCF Transactions on High Performance Computing
- Vol. 6 (3) , 243-262
A survey of compute nodes with 100 TFLOPS and beyond for supercomputers
- 23 May 2024
- Published by Springer Nature in CCF Transactions on High Performance Computing
- Vol. 6 (3) , 243-262
A survey of compute nodes with 100 TFLOPS and beyond for supercomputers
- Open Access
- 1 July 2022
- Published by KSI Research Inc.
- Vol. 2022, 100-105
A Distributed Graph Inference Computation Framework Based on Graph Neural Network Model
- 1 July 2022
- Published by KSI Research Inc.
- Vol. 2022, 100-105
Open AccessA Distributed Graph Inference Computation Framework Based on Graph Neural Network Model
- 1 November 2019
- Published by Elsevier in Information and Software Technology
- Vol. 115, 1-4
Multi-reviewing pull-requests: An exploratory study on GitHub OSS projects
- 1 November 2019
- Published by Elsevier in Information and Software Technology
- Vol. 115, 1-4
Multi-reviewing pull-requests: An exploratory study on GitHub OSS projects