Intel Advanced Vector Extensions 512 (AVX-512) technology acceleration on vector search workloads is now available on OpenSearch 2.17+ domains with 4th generation Intel Xeon Intel instances on the Amazon OpenSearch Service.
C/M/R 7i instances, which use Intel AVX-512, can lead to up to 51% more vector search performance at no additional cost compared to previous R5 Intel instances.
Vector search is used to improve the search quality of applications by encoding content into vectors to find similarities between content and perform tasks.
Intel Xeon Scalable processors on 7th generation Intel instances use AVX-512 to increase the efficiency of vector operations.
Using R7i instances can lead to higher price-performance on OpenSearch vector workloads.
Lucene engine and FAISS engine results show up to a 44% and 51% improvement in price-performance, respectively, when upgrading from R5 to R7i instances.
AVX-512 accelerators contribute to further price-performance gains by reducing power performance costs.
By upgrading to the latest Intel instances, search experiences can be modernized while potentially lowering costs.
Vector search is highly relevant for image to semantic search, recommendations, and generative AI applications.
Vector search offers contextually relevant insights, which can be further enriched by AI for hyper-personalization and integrated with generative models to power chatbots.