Qdrant 1.19.1 Delivers 1.5–2.5x Faster Quantized HNSW Search
qdrant_engine · x · 2026-09-05
Vector database Qdrant released patch version 1.19.1 with substantial under-the-hood performance gains.
- Quantized HNSW search is now 1.5–2.5× faster, using prefetching to saturate memory bandwidth
- Payload-heavy shard transfers are 1.5× faster by using raw payloads
- Other improvements: reworked and batched 4-bit TurboQuant SIMD scoring, batched HNSW searches, batched deletion checks, skipping bad items before priority-queue insertion, and faster L2 cosine preprocessing on AVX
More from Infra
- There's no agreed way to value a GPU running inference—and compute futures now settle on these indexes — AccBalanced · 2026-09-05
- Hermes adds a local backend with Unsloth UD-Q4 quants for DeepSeek-V4-Flash and Qwen models — maximelabonne · 2026-09-05
- AI Now on data center boom: community pushback and 'they won't build them where they live' — AINowInstitute · 2026-09-05
- Gemma 4 Runs 151.4% Faster on Mac via Community MLX Inference Optimization — gajesh · 2026-09-05
- SGLang's Breakable CUDA Graph speeds prefill graph building by 3.8–5.2x — ying11231 · 2026-09-05
- SemiAnalysis: OpenAI's ASIC program is leverage — Altman wins even if the chip loses — MarvinTBaumann · 2026-09-05