Skip to content

Introduction

There are four types of applications currently supported on the Neocortex system, divided into the following individual tracks:

This document is expected to serve as the guiding technical guide for Neocortex users. It is under continuous development. If you have any recommendations for this document, please make sure to share them with the team (see Feedback section).

Neocortex, a system that captures groundbreaking hardware technologies, is designed to accelerate Artificial Intelligence (AI) research in pursuit of science, discovery, and societal good. Neocortex is a highly innovative resource that accelerates AI- powered scientific discovery by vastly shortening the time required for deep learning training, fostering greater integration of artificial deep learning with scientific workflows, and providing revolutionary hardware for the development of more efficient algorithms for artificial intelligence and graph analytics.

Neocortex democratizes access to game-changing compute power otherwise only available to tech giants for students, postdocs, faculty, and others, who require faster turnaround on training to analyze data and integrate AI with simulation. It also inspires the research community to scale their AI-based research and integrate AI advances into their research workflows.

With Neocortex, users are able to apply more accurate models and larger training data, scale model parallelism to unprecedented levels and avoid long wait times when running expensive and time-consuming hyperparameter optimization. The development of new algorithms in machine learning and graph analytics is enabled through this innovative AI platform.

The Neocortex System Slack Organization is now available. Please feel free to join to communicate with other project team members through Slack.

The list below presents the different stages we anticipate every Neocortex research project will go through:

  1. Researcher on-boarding (you will get different emails from neocortex@psc.edu and access to the Neocortex system)
  2. Accessing the Neocortex portal and user documentation (this webpage)
  3. Accessing the Neocortex system (over SSH)
  4. Running a job in the Neocortex system (see the Getting Started Tutorial section for CS-3)
  5. Executing Cerebras reference/sample code (i.e. modelzoo networks, SDK examples, ...)
  6. Transferring your data into the Neocortex environment
  7. Executing your code base in the Neocortex environment
  8. Producing ML key metrics / SDK Detailed project plan
  9. Porting code to Cerebras stack
  10. Optimizing your code for maximum performance