In this video from the GPU Technology Conference, Sumit Gupta from IBM describes how IBM is powering production-level Ai and Machine Learning.
IBM PowerAI provides an on-ramp for enterprise deep learning. PowerAI helped users break deep learning training benchmarks AlexNet and VGGNet thanks to the world’s only CPU-to-GPU NVIDIA NVLink interface. In the next twelve months, new feature development and performance optimizations will advance the future of deep learning, including NVIDIA NVLink 2.0, leaps in distributed training, and tools that make it easier to create deep learning breakthroughs.
In many ways IBM is primed to become a leader in deep learning, having pioneered many of the best practices common in the realm of high-performance computing that are being applied to deep learning. And of course, how we pioneered CPU-to-GPU NVLink in the IBM S822LC best-in-breed server for deep learning. But part of PowerAI’s secret sauce is its deep integration with the open-source deep learning community. Open source communities like OpenPOWER have proven to be dependable sources of innovation, and this was recognized early on. In this vein, Michael detailed how future partnerships will only continue to build the success of PowerAI, bringing in industry leaders like Julia, DL4J, Apache Spark and more.
