Turbovec: Rust vector index fits 10M docs in 4GB and beats FAISS by 3.4x at 4-bit
bibryam · x · 2026-09-21
Turbovec is a local Rust vector index for RAG with Python bindings, built on Google Research's TurboQuant algorithm — a data-oblivious quantizer requiring no separate training phase. Its v1.0 stabilizes the v7 on-disk format; the repo has 17.2k stars.
Key points:
- 2-bit/4-bit compression: a 10M-document corpus drops from 31GB RAM (float32) to 4GB
- Online ingest: add vectors and they're indexed immediately — no train step, no rebuilds as the corpus grows
- Hand-written SIMD kernels (NEON SDOT/SMMLA on ARM, AVX-512 VNNI and vpermb on x86, with AVX2/scalar fallbacks) beat FAISS IndexPQFastScan in every measured config — averaging 3.4x faster at 4-bit and 23% at 2-bit
- Incremental, crash-safe saves via sync(path)
More from Infra
- Agent Substrate Calculator Estimates Session Resume Speed Under Load — pjausovec · 2026-11-12
- Dell expands AI Data Platform: agents have a data-context problem, not a model problem — DavidLinthicum · 2026-11-12
- SGLang Megakernel Pushes Muse Spark to 1,109 tok/s, 2.1x Over Baseline — hsu_byron · 2026-11-12
- Potion UI's local login system turns your PC into a multi-user AI server — cocktailpeanut · 2026-11-12
- AWS Finally Built Its Own AWS Emulator — super_linear · 2026-11-12
- Megakernel is multiprogramming for GPUs: 'GPUs are entering the CPU age' — pranjalssh · 2026-11-12