Research | Jie Zhao

Code Generation and Optimization
- Program transformations using the polyhedral model. Related publications: CC 2018, ACM TACO 2019, MICRO 2020, ACM TOCS 2024.
- Code generation for high-performance architectures. Related publications: ICS 2019, ICPP 2022, PPoPP 2024.
Deep Learning Systems
- Frameworks for deploying deep learning workloads on distributed architectures. Related publications: arXiv 2021.
- Graph compilers that study the fusion of graph nodes. Related publications: MLSys 2022, OSDI 2023.
- Operator compilers that investigate automatic loop transformations and memory managements for tensor operations. Related publications: PLDI 2021, PACT 2022, MLSys 2023.
Numerical Program Analysis
- Floating-point error detection. Related publications: ASE 2023.
- Repairing numerical programs through expression rewriting.

2021.01 - 2024.12
Deep Learning and Tensor Compilers based on the Polyhedral Model
- National Natural Science Foundation of China (Grant No. U20A20226)
- A joint project with the team of Prof. Jidong Zhai and OneFlow Research
- Principal Investigator of the Information Engineering University part
- ¥ 2,600,000 in total; ¥ 800,000 for the Information Engineering University
2019.01 - 2021.12
Analysis and Optimization of the Precision of Mathematical Functions on Domestic Processors
- National Natural Science Foundation of China (Grant No. 61802434)
- Researcher (Principal Investigator: Prof. Jinchen Xu)
- ¥ 250,000
2018.01 - 2020.12
Polyhedral Compilation Techniques for Heterogeneous Architectures
- National Natural Science Foundation of China (Grant No. 61702546)
- Principal Investigator
- ¥ 240,000