Data Crammed into few Bits with Random Access
Data structures often waste considerable amounts memory in order to support faster operations, such as hash tables operating at a load factor of 50%. Sometimes this is okay. We are concerned with the case when it’s not.
Our Goals
We develop data structures that combine fast execution speeds and excellent memory efficiency. We pursue both theoretical understanding of the underlying mathematical structures as well as highly engineered implementations that excell in relevant benchmarks.
Topics
Group members have considerable expertise in the following topics. We intend to keep pushing for further improvements in those and related areas.
- Perfect Hashing
- Retrieval Data Structures
- Filter Data Structures
- Cuckoo Hashing
- Set-Sketches
- Relaxed (FIFO/Priority) Queues