GigaSpaces, Intel partnership makes AI accessible without expensive hardware or GPUs Posted on : Sep 23 - 2017

A new partnership between GigaSpaces and Intel aims to simplify AI deployments in the enterprise, eliminating the need for expensive specialized hardware and reducing reliance on GPUs, the company announced in a press release.

The new solution will be built on an integration between GigaSpaces' InsightEdge platform and the Intel BigDL open source deep learning library for Apache Spark, the release said. The solution brings a distributed deep learning framework the drives enterprise insights, according to a GigaSpaces blog post.

GigaSpaces, which provides in-memory computing (IMC) platforms, has worked in industries like financial services, healthcare, transportation, and retail, and noted a rise of AI adoption in those fields, the release said. As such, the firm has boosted its own analytics offerings through Apache Spark, now integrating with BigDL to further broaden its portfolio.

"The BigDL and AI portfolio provide an infrastructure-optimized solution for deep learning workloads leveraging Intel Xeon Scalable processors," the release said. "Together the technologies fill a critical market gap by creating an intelligent insight platform that makes it easy to innovate on real-time advanced analytics applications with low risk and TCO."

The new integrated offering offers three key benefits: Cost savings, simplicity, and scalability.

Because Intel's BigDL makes expensive deep learning hardware unnecessary, companies can use low-cost compute infrastructure instead. This way, they can save money when running deep learning workloads, as they don't have to put up the cash to invest in GPU performance, the release noted.

To say deep learning is complex is an understatement, but the InsightEdge analytics stack, leveraging BigDL and Apache Spark, minimizes the number of moving parts involved with training deep learning workflows and simplifies the process, the release said.

The partnership will also improve scalability, as users can scale workloads like text mining, image recognition, and more "from a handful of machines to thousands of nodes in the cloud or on-premises, using the same application assets and deployment lifecycle," the release noted. View More