State of the Art Hardware Meets State of the Art Software

The rapid advance of AI technologies has driven a fresh wave of innovation in computational hardware designed to enable deep learning models with millions or billions of parameters and massive amounts of data.  

And as companies across every industry have invested in applying AI to their most important problems they’ve needed new tools and new infrastructure to put that power in the hands of their Data Science and ML Teams.

Graphcore is partnering with Spell to bring together the most advanced deep learning computation and infrastructure software in an integrated offering that is powerful, easy to use, and highly cost-effective. And we’re making it free to try.

Designed for AI Compute

tunnel with cogwheel in center

Designed for AI Compute

The Graphcore Intelligence Processing Unit (IPU) is a completely new processor, specifically designed for AI compute. The IPU’s unique architecture lets AI researchers undertake entirely new types of work, not possible using current technologies, to drive the next advances in machine intelligence.

Hardware-agnostic Infrastructure Management

server fans

Hardware-agnostic Infrastructure Management

Spell’s unique, hardware-agnostic AI Infrastructure management software hides the underlying hardware environment from the end-user, eliminating the need for AI engineers to develop custom, hardware-specific code and workflows for managing and optimizing compute and data management  resources.

Seamless Graphcore Integration

graphcore arrow

Seamless Graphcore Integration

The seamless integration of Graphcore’s innovative AI acceleration hardware with Spell’s powerful deep learning orchestration and automation software dramatically reduces training processor cost and greatly simplifies lifecycle management for the increasingly large and complex models for Natural Language Processing, Image Recognition, Interactive Voice Response, and other advanced AI applications.

Take a Free Test Drive

Sign up for a free Spell Community account, select the Graphcore IPU option, and start building and training your deep learning model.  

Your Spell Community account also gives you free access to NVIDIA GPUs, allowing seamless access to the right performance and resource efficiency for any given workload.  Just `spell run` your experiment up to an automated cluster, all through a single, simple UI and command line environment.