Sirius: A GPU-Native SQL Engine (Xiangyao Yu)
- Speaker:
- Xiangyao Yu
- Date:
- Mon Apr 27, 2026 @ 04:30pm EDT
- Date:
- Mon Apr 27, 2026
- Time:
- 04:30pm EDT
- Location:
- https://cmu.zoom.us/j/99830697483?pwd=RLKiHNDLPvOyoCMHSBGmWfbba4ZAb4.1Zoom
- Title:
- Sirius: A GPU-Native SQL Engine
- System:
- Sirius
- Video:
- YouTube
Talk Info:
GPUs have evolved into powerful, cost-efficient engines for general-purpose parallel compute. This talk introduces Sirius, a GPU-native SQL engine developed in partnership between UW-Madison and NVIDIA. Sirius enables drop-in GPU acceleration for DuckDB and other SQL databases, without changing the user interface. By treating the GPU as the primary execution engine, Sirius achieves strong performance by leveraging the massive compute power and memory bandwidth of modern GPUs. On TPC-H at 1TB scale, Sirius delivers a 9× improvement in performance per dollar compared to DuckDB on CPU, unlocking faster and more cost-efficient interactive analytics. We will discuss how Sirius works under the hood, the hardware and software trends making GPU analytics practical today, and why GPUs represent the future of analytical processing.
This talk is part of the PostgreSQL vs. The World Seminar Series.
Bio:
Xiangyao Yu is an Assistant Professor at the University of Wisconsin-Madison working on database systems. He is a recipient of the NSF CAREER Award, the Sloan Research Fellowship, and the VLDB Early Career Research Contribution Award.