Samyama: A Unified Graph-Vector Database with In-Database Optimization, Agentic Enrichment, and Hardware Acceleration
Madhulatha Mandarapu, Sandeep Kunkunuru · 2026
Read Samyama: A Unified Graph-Vector Database with In-Database Optimization, Agentic Enrichment, and Hardware Acceleration by Madhulatha Mandarapu, Sandeep Kunkunuru (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-08036.txt.
Description
Mandarapu and colleagues present Samyama, a graph-vector database in Rust that joins workloads usually split across separate engines: graph storage, vector search, analytics, and optimization solvers. It pairs a RocksDB-backed store with a vectorized executor, a cost-based planner, and a CSR analytics engine. It also offers RDF and SPARQL support, 22 metaheuristic solvers, HNSW indexing, and agentic enrichment that grows the graph with language models. A revised version withdraws earlier claims as unreproducible, including an 8.2x GPU speedup, and re-measures the system on rented Linux hosts.
Catalog record
| CREATOR | Madhulatha Mandarapu, Sandeep Kunkunuru |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26053 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.08036