IBM Watson gains processing boost with Nvidia and Power8 hardware

OpenPower helps cloud-powered cognitive computing service carry out big data analytics faster

IBM has given its Watson cognitive computing platform an injection of processing power to speed up the way it carries out analytics on big data.

Watson's natural language comprehension, deep learning and advanced analytics capabilities are delivered through IBM's cloud platform, but the underpinning infrastructure is very much a hardware affair.

IBM has now integrated Nvidia's Tesla accelerated graphics platform into Watson's hardware foundations in the form of Tesla K80 GPUs coupled with IBM's Power8 processors commonly used to power servers that support enterprise and scalable cloud infrastructure.

This injection of power has given Watson the ability to carry out rank and retrieve actions on data accessed through APIs at 1.7 times its normal speed, according to IBM, while tapping into Nivida's GPU's gives Watson a 10-fold hike in processing power.

The boost means that customers accessing Watson through the cloud should see faster natural language processing queries against data, particularly unstructured data sets, providing faster answers to big data questions and a greater potential to carry out actions on data-driven insights.

IBM Watson chief technology officer Rob High used the example of a call centre to show how Watson's power boost can benefit customers.

"If a call centre agent is responding to an individual's health and insurance query, the agent will be able to leverage Watson's natural language processing technology to obtain an answer in real time even faster and cheaper than before," he said.

High also explained that the combination of Power8 processor architecture and the Tesla platform enhances Watson's deep learning capabilities and allows the cognitive computing system to reason deeply about visual information and speech.

Bolstering Watson's power is all part of IBM's plan to build out the analytics service into a cognitive computing force that can be deployed across a wide range of industries, from predictive maintenance in manufacturing to helping athletes perform harder for longer.

Much of the power benefits to Watson come from the OpenPower Foundation, a development organisation set up by IBM, Google, Nvidia, Mellanox and Tyan to explore how the Power architecture can be built on to facilitate the creation of new hardware and software products for use in data centres.

The Foundation also helps to propel the Watson ecosystem of developers and startups working with APIs that allow its cognitive computing capabilities to be embedded into apps and analytics services, according to High.

Lauri Saft, vice president of Watson Ecosystem at IBM, pointed out that much of Watson's future lies in nurturing a healthy ecosystem of users and developers who can find new ways to put Watson to use that IBM cannot do alone.

"When we look at the growth of Watson, the idea is how we can get this into the hands of everybody no matter their experience, be it in your doctor's office, training with your personal trainer, or with the retail associate you are interacting with," she told V3 during an interview at IBM Insight 2015.

That being said, IBM did create Watson Health under its own steam, aided with two startup acquisitions, to provide healthcare workers with an open platform to share and make better use of medical research and health data.