Qdrant's Filterable HNSW Outperforms ACORN in Vector Search Benchmarks
qdrant_engine · x · 2026-08-19
Qdrant compared its Filterable HNSW algorithm against ACORN for handling filtered vector search. While ACORN works around disconnected graph islands at query time, Qdrant adds edges between points sharing indexed payload values to maintain connectivity.
Benchmarks on 1M vectors with a 1% filter show Filterable HNSW achieving 99.8% recall at 1.0ms, significantly outperforming ACORN's 67.7% recall at 4.7ms. ACORN remains useful for broad values and AND filters, but performs best when supported by a filter-built graph.
More from Infra
- DeepSeek-V4-Flash benchmarks: 286 tok/s single-stream, +34% boost with speculative decoding — mcraddock · 2026-08-19
- Initialization-Free Bundle Adjustment Revisited: A Controlled Study — ssh4net · 2026-08-19
- Are old datacenter cards like P40 or MI50 still worth it for local AI? — Far-Classic-9963 · 2026-08-19
- Window Assassin: Tray tool to kill processes hogging 1+ GB of VRAM — b2kdaman · 2026-08-19
- Public underestimates physical infrastructure needed for the cloud — csuwildcat · 2026-08-19
- China's National Supercomputing Network offers tokens 8x cheaper than DeepSeek — teortaxesTex · 2026-08-19