Feature spotlights

Enables rapid deployment

Includes most popular deep learning frameworks, including all required dependencies and files, precompiled and ready to deploy. The entire AI suite has been validated and optimized to run reliably on accelerated Power servers.

Offers simplified maintenance

The entire software stack is validated with each new framework release, patch, or security update. Patches & updates arrive via repository. Knowledgeable IBM support teams are available for assistance.

Supports larger, more accurate models

Large model support facilitates the use of system memory with little to no performance impact, yielding significantly larger and more accurate deep learning models.

Delivers faster insights

Leverages unique capabilities of accelerated Power servers, delivering performance unattainable on commodity servers, Provides for hyper-parameter search and optimization, and elastic training to allocate the resources needed to optimize performance; and Distributed deep Learning provides for rapid insights at massive scale.

Provides enterprise multi-tenancy and dependability

Supports multiple users and lines of business with Multi-tenancy, end-to-end security, including role-based access controls. Provides for centralized management and monitoring of the infrastructure and workflows.

Protected by IBM support

Watson Machine Learning Accelerator software and the accelerated Power servers it runs on are supported by knowledgeable IBM technical support.

Efficient and cost-effective

Watson Machine Learning Accelerator fully exploits the efficiency of accelerated IBM Power servers and provides a broad range of capabilities, enabling it to deliver faster and more accurate insights from AI with less infrastructure.

Provides flexibility to run additional applications

Watson Machine Learning Accelerator runs on IBM Power accelerated servers HPC, a platform that runs not only your deep learning but also a wide variety of HPC and High Performance Data Analytics workloads.

Incorporates the most popular deep learning frameworks

Watson Machine Learning Acclerator gives access to Power-optimized versions of all of the most popular deep learning frameworks currently available, including TensorFlow, Caffe, and PyTorch.

Technical details

Software requirements

Red Hat Enterprise Linux® 7.5 and requisite third-party software from NVIDIA

    Hardware requirements

    Watson Machine Learning Accelerator is optimized for the IBM Power System AC922 and the IBM Power System S822LC for High Performance Computing.

    • 2 IBM POWER9 or POWER8 CPUs
    • 128GB of memory, or greater, recommended
    • NVIDIA® Tesla® P100 or V100 with NVLink GPUs are required
    • NVIDIA NVLink interface to Tesla GPUs strongly recommended