Xu Dai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Xu Dai has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Shouyi Yin
Tsinghua University
3 shared publications - Yang Hu
Tsinghua University
3 shared publications - Zheng Xu
Tsinghua University
3 shared publications - Dehao Kong
Tsinghua University
2 shared publications - Jiaxin Liu
Tsinghua University
2 shared publications - Shaojun Wei
Tsinghua University
2 shared publications
Latest Publications
- 24 August 2026
- Published by Springer Nature
- p. 579-585
WaferSim: A Simulation Infrastructure for LLM Service on Wafer-Scale Chips
- 24 August 2026
- Published by Springer Nature
- p. 579-585
WaferSim: A Simulation Infrastructure for LLM Service on Wafer-Scale Chips
- Top Cited PapersOpen Access
- 20 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-17
WSC-LLM: Efficient LLM Service and Architecture Co-exploration for Wafer-scale Chips
- 20 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Top Cited PapersOpen AccessWSC-LLM: Efficient LLM Service and Architecture Co-exploration for Wafer-scale Chips
- Open Access
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
GSPO: A Graph Substitution and Parallelization Joint Optimization Framework for DNN Inference
This work proposes GSPO, an automatic unified framework that jointly applies graph substitution and parallelization for DNN inference. GSPO uses a joint optimization computation graph (JOCG) to represent graph substitution and parallelization at the operator level. Then, a novel cost model customized for joint optimization is used to evaluate the computation graph execution time quickly. With the graph partition and backtracking search algorithm, GSPO can find the optimal joint optimization solution within an acceptable search time. Compared to existing frameworks applying graph substitution or parallelization, GSPO can achieve up to 27.1% end-to-end performance improvement and reduce search time by up to 94.3%.
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessGSPO: A Graph Substitution and Parallelization Joint Optimization Framework for DNN Inference
This work proposes GSPO, an automatic unified framework that jointly applies graph substitution and parallelization for DNN inference. GSPO uses a joint optimization computation graph (JOCG) to represent graph substitution and parallelization at the operator level. Then, a novel cost model customized for joint optimization is used to evaluate the computation graph execution time quickly. With the graph partition and backtracking search algorithm, GSPO can find the optimal joint optimization solution within an acceptable search time. Compared to existing frameworks applying graph substitution or parallelization, GSPO can achieve up to 27.1% end-to-end performance improvement and reduce search time by up to 94.3%.