Columbia

Architecture and Design

Lab

DPU: Data Processing Unit

This research project examines the design and deployment of specialized processors to improve the throughput and energy efficiency of large-scale data processing. In particular, our focus is on relational database query processing and the design of a database processing unit or DPU. Like GPUs, DPUs are domain-specific processors which can support a range of database oriented computations. Our first DPU design focuses on analytic queries of large data sets.

Publications

Q100: The Architecture and Design of a Database Processing Unit, L. Wu, A. Lottarini, T. Paine, M. A. Kim, K. A. Ross, In ASPLOS , 2014.

pdf
slides

Navigating Big Data with High-Throughput, Energy-Efficient Data Partitioning, L. Wu, R. J. Barker, M. A. Kim, K. A. Ross, In ISCA , 2013. Top Picks in Computer Architecture Selection.

pdf
slides

Support

This material is based upon work supported by the National Science Foundation under Grant No. CCF-1065338. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.