GrAND: GPU-based Dynamic Graph Indexes for Approximate Nearest Neighbour Search
What it is
GrAND is a set of GPU-native algorithms that enable dynamic updates (insertions and deletions) in graph-based Approximate Nearest Neighbour Search (ANNS) indexes. It optimizes high-throughput search by consolidating graph repair across batches and using a lock-free find-and-replace strategy for parallel updates.
The problem it solves
Existing GPU graph indexes are often static or suffer from high computational overhead and memory interference when handling frequent data updates.
What you could use it for
- Real-time vector database updates for streaming data
- High-throughput ANN search with frequent insertions/deletions
- Large-scale similarity search in dynamic datasets
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.