IBM Releases POWER9, Enhancing Performance Across AI Frameworks

IBM is launching its next-generation Power Systems Servers incorporating its newly designed POWER9 processor, advancing performance improvements across popular AI frameworks.

Built specifically for compute-intensive AI workloads, the new POWER9 systems are capable of improving the training times of deep learning frameworks by nearly which allows enterprises to build more accurate AI applications.

“We essentially have three interfaces that enable this data movement to this super highway,” said Sumit Gupta, machine/deep learning product and Business Leader at IBM.

The new POWER9-based AC922 Power Systems are the first to embed technologies PCI-Express 4.0, next-generation NVIDIA NVlink 2.0, and OpenCAPI which combined can accelerate data movement.

The system was designed to drive demonstrable performance improvements across popular AI frameworks such as Chainer, TensorFlow and Caffe, as well as accelerated databases such as Kinetica.

As a result, data scientists can build applications faster, ranging from deep learning insights in scientific research, real-time fraud detection and credit risk analysis.

AI, machine learning, deep learning, data analysts and high performance computing users will benefit the most from this update, Gupta said.

“The toughest workloads out there and the fastest growing ones, those are the places where we are seeing the maximum benefit,” Gupta said.

Deep learning is a fast growing machine learning method that extracts information by crunching through millions of processes and data to detect and rank the most important aspects of the data.

To meet these growing industry demands, four year ago, IBM set out to design the POWER9 chip on a blank sheet to build a new architecture to manage free-flowing data, streaming sensors and algorithms for data-intensive AI and deep learning workloads on Linux.

IBM serves a community of over 300 OpenPOWER Foundation and OpenCAPI Consortium members. This includes SuperMicro, Inspur, Xilinx, Nutanix, NVIDIA, Inventec, Wistron, Mellanox Technologies, Amphenol, Rackspace, Molex, Alpha-Data, Toshiba, Rambus, MicroSemi, Google, Tektronics and Western Digital -- many of whom are also expected to launch products and solutions for POWER9.

“We’ll continue to enhance the performance advantage on AI workloads which includes databases, data stores, Hadoop, machine learning, and data analytics,” Gupta said.

For more information about this update, visit