The Secret to Jumpstarting Your AI Strategy

Organizations across all industries are seeing the value and competitive advantage of having an artificial intelligence (AI) strategy.  AI applications include consumer sentiment analysis, market prediction and fraud risk, drug discovery and genomics, video classification, medical imaging, language processing, and sensor analysis, and numerous other AI use cases.

The challenges of implementing an AI strategy are many. While various server vendors offer a single AI system, such solutions simply don’t scale past the single user they were designed for, though meant for use by a team of data scientists.  Some solutions require enough servers to start filling a rack in order to achieve the computer processing needed to create more powerful AI models. Finally, perhaps the biggest obstacle faced when adopting an AI project is the time it takes to configure the server, prepare the data, build and train the model, and deploy and infer for deep learning. Often it takes weeks and months to get an AI infrastructure up and running.

The solution to these challenges and jumpstarting your AI project?

IBM recently released an AI Starter Kit that provides an entry point for clients starting an AI environment. Equipped with IBM’s Deep Learning frameworks bundle, Watson Machine Learning Accelerator (WMLA), formerly called PowerAI, the solution is designed for use by 3-5 data scientists. The kit helps facilitate clients’ AI adoption journey from experimentation to production. More servers can also be added later for an enterprise-scale solution.

The kit is a ready-to-go offering, pre-built and installed in manufacturing, including the setup of all components and the installation and configuration of all software. A complete system is shipped to clients and deployed on-site, ready to use. To make it even easier, IBM Lab Services is included to simplify set-up and perform a knowledge transfer with clients.

 Other AI Starter Kit Features

  • 2 IBM Power Systems Accelerated Compute AC922 Servers
  • 2 POWER9 CPUs
  • 4 NVIDIA TESLA V100 GPUs (32 GB)
  • 512 GB memory, 3.8 TB storage, and RHEL 7.6
  • A storage rich IBM LC922 server with 128 GB memory and 7 X 10 TB drives with NFS/Scale for temporary node storage
  • 10 GbE or 100 GbE used to connect nodes

IBM Power Systems AI Starter Kit was designed to be a game changer for client AI projects. The combination of Watson Machine Learning Accelerator and AC922 servers reduces AI deployment times from weeks and months, down to hours.  These are the same AC922 servers that make up the world’s fastest super computer, Summit.

Use Cases

Sirius has unmatched expertise in IBM Power. An IBM Platinum Business Partner and the largest IBM solution provider in the United States, Sirius has helped clients deploy IBM Power AC922 servers with WMLA for their AI projects. A couple of use cases where WMLA on Power AC922 provided an ideal solution include medical imaging and radiology for cancer research, and predictive pipeline analytics for a utility company.

Sirius also has a partnership with NVIDIA. With comprehensive, wide-range capabilities, Sirius experts have helped our clients deploy IBM Power AI technology in multiple industries, including academic, utility and medical environments, using both Power8 and Power9 servers. Contact a Sirius expert for more details today.

By |2019-08-20T11:34:49-06:00August 21st, 2019|Blog|Comments Off on The Secret to Jumpstarting Your AI Strategy

About the Author:

Niel Balsino is a Senior Power Engineer-Linux with Sirius Computer Solutions.