Nvidia and IBM team up for PowerAI deep learning software push

Nvidia and IBM have announced a deep learning partnership dubbed PowerAI as both companies seek to advance their efforts in this field.

The software, which is designed to run on IBM's Linux-based S822LC HPC (high performance computing) server, can handle GPU-based workloads in AI and deep learning areas by combining custom versions of three existing popular frameworks.

Torch, Theano and Caffe, as Linux-based software, are all optimised to run on the S822LC.

"PowerAI democratises deep learning and other advanced analytic technologies by giving enterprise data scientists and research scientists an easy-to-deploy platform to rapidly advance their journey on AI," said Ken King, general manager of IBM OpenPower.

"Coupled with our HPC servers built for AI, IBM provides what we believe is the best platform for enterprises building AI-based software, whether it's chatbots for customer engagement or real-time analysis of social media data."

IBM stated in the release notes that the technology can be used for a broad set of purposes.

"For example, new driver assist technologies rely on machine and deep learning patterns to recognise objects in a rapidly changing environment, and personal digital assistant technology is learning to categorise information contained in emails and text messages based on context," said IBM.

"And in the enterprise, machine and deep learning applications can be used to identify high-value sales opportunities, provide assistance in call centres, detect instances of intrusion or fraud and suggest solutions to technical or business problems."

PowerAI is free of charge to organisations that have already adopted S822LC hardware, which IBM claimed can work up to twice as fast as rival servers. The firm reckons that S822LC will run even faster with the addition of PowerAI.

"Our innovation with IBM on Nvidia NVLink [NVidia's communications protocol between CPU and GPU] has created new opportunities for Power in the deep learning and analytics market," said Nvidia vice president of accelerated computing Ian Buck.

"Nvidia's GPUDL libraries in PowerAI will provide world-class high-performance tools to power GPU-accelerated deep learning applications."