GSI Technology's k-NN Plug-in for OpenSearch

GSI is proud to participate with OpenSearch as a partner with other leading community contributors. OpenSearch makes it easy to ingest, search, visualize, and analyze your data.

GSI Technology is the developer of the breakthrough Associative Processing Unit (APU) that allows high performance K and Approximate Nearest Neighbor (KNN and ANN) implementations and scaling similarity searches.

Our contribution to the OpenSearch community is underlying APU hardware acceleration technology as a simple acceleration-as-a-service plug-in for search algorithms run in AWS.

Our service allows OpenSearch users to dramatically accelerate searches, add dense_vector support, and multimodal queries to large databases, including very large billion scale ones. This increases performance and drives down power consumption, providing lower cost per query to the OpenSearch community.

The GSI APU can accelerate and provide additional capabilities to your OpenSearch workloads with a simple plug-in to your AWS workflow. Contact us to request a free trial.

Gemini APU

GSI’s Gemini APU is an in-memory Associative Processing Unit that can eliminate von Neumann data transfer inefficiency and significantly improve large database similarity searches, high null-data database searches, and fast vector searches. The APU is used in applications such as e-commerce object recognition, bio and chem-informatics, cybersecurity, Natural Language Processing (NLP), and visual search just to name a few.

GSI Technology Partners with AWS OpenSearch

GSI Technology contributes to the OpenSearch community by providing the underlying APU hardware acceleration technology for search algorithms in AWS and on premises, allowing OpenSearch users to accelerate searches, add dense_vector support, and multimodal queries to large databases, including very large billion scale. This increases performance and drives down power consumption, providing lower cost per query to the OpenSearch community.

Additional APU Use Cases

GSI Technology Takes on DSP with Deep Learning

The GSI Leda-G® PCIe board leverages the GSI APU to improve existing algorithms in spectrum management and monitoring in both the commercial and aerospace industry providing innovative approaches to HPC requirements in next generation communication networks.

Molecular Search with the Gemini APU

Gemini is a patented processing technology featuring massive parallel data processing, compute and search; inplace, directly in the memory array. This architecture gives the APU a performance edge in the acceleration of similarity search applications.

Facial Recognition with the Gemini APU

Gemini® delivers accurate search results while reducing search times in large databases from many minutes to fractions of a second, all with very low power consumption.