Skip to content

Releases: andrestubbe/FastAIVectorDB

FastAIVectorDB 0.1.1

Choose a tag to compare

@andrestubbe andrestubbe released this 08 Aug 10:11

Release 0.1.1 featuring native C++ SIMD similarity search, JMH benchmarks, and standardized documentation suite.

FastAIVectorDB 0.1.0

Choose a tag to compare

@andrestubbe andrestubbe released this 11 Jul 02:53

================================================================================
YOUTUBE RELEASE: FastAIVectorDB v0.1.0

TITLE:
JVM Vector Search in <1ms: FastAIVectorDB v0.1.0 Native Release! 🚀

DESCRIPTION:
FastAIVectorDB v0.1.0 is live! FastAIVectorDB brings raw C++ execution speed to Java vector similarity searches via optimized JNI bindings, coupled with a complete, thread-safe pure Java fallback.

Designed to bypass the garbage collector and eliminate external DB dependencies, it delivers sub-millisecond similarity scans directly inside your Java process.

KEY HIGHLIGHTS:

  • [Native Engine]: Highly optimized C++ linear similarity scan compiled natively for x64 Windows.
  • [Java Fallback]: Automatically switches to a pure-Java memory store if the native DLL is not loaded.
  • [JNI Bridge]: Zero-allocation bridge bindings to query indices without GC stalls.

RESULTS:

  • Native similarity lookup: <1 ms
  • Java Fallback lookup: ~5 ms
  • Speedup: ~5-10x faster execution

LINKS:
GitHub FastAIVectorDB: https://github.com/andrestubbe/FastAIVectorDB
FastJava Ecosystem: https://github.com/andrestubbe

Part of the FastJava Ecosystem — Making the JVM faster.

Made with ⚡ by Andre Stubbe

TAGS:
Java, Performance, Native, C++, JNI, Windows, Optimization, VectorDatabase, MachineLearning, Embeddings, AI, DeveloperTools