Edit: Some links no longer work
Originally posted May 28, 2019 on AIXchange
Following up on last week’s AI-themed post, I encourage you to check the extensive PowerAI documentation available from the IBM Knowledge Center.
You’ll find instructions on planning, installing frameworks and PowerAI system setup, along with frequently asked questions, a developer portal, and more.
There are also two new Redbooks that get into more concepts and information around Deep Learning and AI and Big Data.
Here are a couple short excerpts from the deep learning Redbook. First, from page 22 section 2.1: “What is IBM PowerAI?”
IBM PowerAI is a package of software distributions for many of the major deep learning (DL) software frameworks for model training, such as TensorFlow, Caffe, Chainer, Torch, and Theano, and their associated libraries, such as CUDA Deep Neural Network (cuDNN), and nvCaffe. They are extensions that take advantage of accelerators, for example, nvCaffe is NVIDIA extension to Caffe so that it can work on graphical processing units (GPU). As with nvCaffe, IBM has an own extension to Caffe, which is called IBM Caffe. Furthermore, the IBM PowerAI solution is optimized for performance by using the NVLink-based IBM POWER8 server, the IBM Power S822LC for High Performance Computing server, and its successor, the IBM Power System AC922 for High Performance Computing server. The stack also comes with supporting libraries, such as Deep Learning GPU Training System (DIGITS), OpenBLAS, Bazel, and NVIDIA Collective Communications Library (NCCL).
Here’s more from section 2.2:
IBM PowerAI provides the following benefits:
Fast time to deploy a DL environment so that clients can get to work immediately:
- Simplified installation in usually less than 1 hour
- Precompiled DL libraries, including all required files
Optimized performance so users can capture value sooner:
- Built for IBM Power Systems servers with NVLink CPUs and NVIDIA GPUs, delivering performance unattainable elsewhere
- Distributed DL, taking advantage of parallel processing
Designed for enterprise deployments:
- Multitenancy supporting multiple users and lines of business (LOBs)
- Centralized management and monitoring by integrations with other software
IBM service and support for the entire solution, including the open source DL frameworks.
One thing I can tell you from experience: the most recent releases of PowerAI are much easier to install than the earlier versions. And upgrading is simple enough. For instance, I was working on Redhat with an older PowerAI version, so I followed this information to upgrade it:
Taking those steps, in that order, we were able to start working with the latest version of PowerAI.