SAN JOSE, CA — Iguazio, supplier of the high-performance platform for serverless and machine studying purposes, has introduced native integration with NVIDIA GPUs to eradicate knowledge bottlenecks, present higher scalability and shorten time to manufacturing. Iguazio’s platform powers machine studying and knowledge science over Kubernetes, enabling automated scaling to a number of NVIDIA GPU servers and fast processing of a whole lot of terabytes of knowledge.
Iguazio’s knowledge science platform supplies:
Serverless capabilities that run data-intensive purposes on NVIDIA GPU-enabled servers.
Integration of its database with RAPIDS, NVIDIA’s open-source machine studying libraries, for sooner and scalable knowledge processing.
The usage of Iguazio’s serverless capabilities (Nuclio) improves GPU utilization and sharing, leading to virtually 4 occasions sooner utility efficiency when in comparison with using GPUs inside monolithic architectures. Nuclio is fifty occasions sooner than serverless options that don’t provide GPU assist, corresponding to Amazon’s Lambda. Serverless and Kubernetes goal key challenges in knowledge science: they simplify operationalization, eradicate guide devops processes and minimize time to market.
Samsung SDS will use Iguazio to hurry up machine studying purposes and leverage the automated scaling of Iguazio’s serverless framework to extend effectivity and sharing. Samsung SDS introduced its funding in Iguazio on March sixth. The corporate has accelerated its pipeline with Iguazio to streamline the supply of clever purposes, analyzing fashions constructed immediately in its manufacturing surroundings and producing predictions.
“The mixing of Iguazio with NVIDIA RAPIDS supplies a breakthrough in efficiency and scalability for knowledge evaluation and a broad set of machine studying algorithms,” stated Iguazio CTO Yaron Haviv. “Our platform is already powering a collaborative surroundings and driving cross-team productiveness for the processing of huge quantities of knowledge and parallel computing.”
“The RAPIDS suite of open-source libraries allows scaling of GPU-accelerated knowledge processing and machine studying to multi-node, multi-GPU deployments,” stated Jeffrey Tseng, Director of Product, AI Infrastructure at NVIDIA. “Iguazio enhances the compute energy of GPUs by immediately connecting GPUs to large-scale, shared knowledge infrastructure.”