Operate Machine Learning Models at Scale

A proven cloud portability solution that enables machine learning and inference from edge to core

Drawing Insights from Data Requires a Robust Training and Model Deployment Platform

As AI infrastructure evolves, it becomes more and more dependent on high performance computing resources to improve performance. The Deep Neural Networks (DNNs) currently being built and trained are more complex than ever before. These are the DNNs that get famous for outclassing humans at tasks like image recognition and beating world class Go players. Training these models is a hugely compute-intensive process, so much can be gained by reducing training times with optimal hardware and software platforms.

Penguin Computing provides specialized hardware and software technologies that accelerate DNN training and orchestrate DNN container deployment. We partner with Red Hat to deploy DNNs in containers using the Red Hat OpenShift Container Platform. Once a DNN model has been well trained and is ready for deployment, it can be deployed on a myriad of different hardware platforms around the world using Openshift. As the new container arrives on the edge inferencing system, it is quickly loaded and begins making new inferences on the data it receives. As the deployed DNN model sends inferenced data back to the training platform, the model can be refined and optimized, allowing for constant improvement in inferencing models. Using a container based system, this management of DNN models becomes easy, since you can update and roll back models as needed across hundreds or thousands of devices anywhere.


The combination of Penguin Computing hardware and Red Hat® OpenShift Container Platform 4 provides an engineered and optimized artificial intelligence/machine learning (AI/ML) infrastructure to run and monitor DNNs in Red Hat® Enterprise Linux® containers. Along with excellent performance, this architecture provides evidence of how OpenShift provides easy access to high-performance, containerized machine learning training on Penguin Computing hardware.

Read the white papers today to see how Penguin Computing® Cloud Orchestrator with Red Hat® OpenShift® 4 can empower your organization with either open compute 21-inch Linux servers or traditional 19” Linux servers.

gpu accelerated computing penguin

GPU-accelerated computing can help you:

  • Improve lease bidding
  • Increase the ROI on exploration
  • Optimize drilling effectiveness
  • Get faster time-to-market with new additive, blends, and other products
  • Make faster emergency planning and emergency management decisions