Zilliz introduces Milvus 2.4 with GPU indexing support for CAGRA

Zilliz, an innovator in vector database technology, is proud to announce the release of Milvus 2.4, which sets a new standard in vector search capabilities by revolutionizing GPU indexing with the CUDA-accelerated NVIDIA Graphics Index for Vector Retrieval (CAGRA), part of the RAPIDS cuVS library.

Building on Zilliz's commitment to innovation, Milvus 2.4 extends the GPU indexing capabilities from the introduction of the GPU IVF-Flat and GPU IVF-PQ indexes last year. GPU indexing represents a significant milestone in vector database technology, allowing Milvus 2.4 to further outpace traditional CPU-based indexes such as HNSW. By harnessing the power of GPU acceleration, Milvus 2.4 delivers noticeable performance gains, especially when working with large data sets, providing lightning-fast search responses and unprecedented efficiency for developers.

Accelerated vector search capabilities are key to improving workflows with advanced search. Customers can use Milvus with NVIDIA NeMo Retriever microservices announced at NVIDIA GTC, the worldwide AI conference.

"We are pleased to introduce next-generation GPU indexing in Milvus 2.4, a release that underscores our commitment to advancing GPU-enabled vector database technology," said Charles Xie, CEO of Zilliz. "These features provide developers with unprecedented capabilities, enabling them to build highly efficient and scalable applications that utilize a combination of vector search and GPUs."

"Vector databases play a key role in the performance of AI models," said John Zedlewski, senior director of accelerated data science at NVIDIA. "GPU acceleration in Milvus 2.4 with cuVS will improve a wide range of generative AI applications built on vector search, including NVIDIA NeMo Retriever."

In addition to GPU indexing, Milvus 2.4 now supports GPU-based brute force search, further improving search performance without compromising speed. This combination of features allows developers to create highly efficient and scalable applications that utilize the full potential of vector search technology.

Milvus 2.4 is now available for download. Check out the latest features and enhancements at the NVIDIA GTC session titled: Perform high-performance searches, improve data freshness, and increase recall accuracy with GPU-accelerated vector search and RAG workflows or on the Milvus website .

Zilliz is a leading vector database company founded by the engineers who created Milvus, the world's most widely used open source vector database. Zilliz next-generation database technologies help organizations rapidly build AI/ML applications and unlock the potential of unstructured data. By simplifying the management of complex data infrastructure, Zilliz aims to bring the power of artificial intelligence to every corporation, organization and individual.

Zilliz is headquartered in Redwood Shores, California. The company is backed by high-profile investors such as Aramco's Prosperity7 Ventures, Temasek's Pavilion Capital, Hillhouse Capital, 5Y Capital, Yunqi Partners, Trustbridge Partners and others. Zilliz's technologies and products help more than 5,000+ organizations around the world easily build artificial intelligence applications across a variety of use cases.