Penguin Computing Artificial Intelligence Practice Overview
From government labs to the premier artificial intelligence (AI) lab in the private sector, Penguin Computing has been designing, building, and running AI systems for years. In fact, we’ve built so many AI systems that we won the 2017 NVIDIA Americas Partner of the Year award and have grown our business with the graphics processing unit (GPU) leader by 778% over the past four years. But we know that all AI systems are not the same. They can’t be, because every organization has different goals and constraints. That’s why Penguin Computing brings together experts in all fields to create balanced AI systems.
What capabilities does the AI Practice have?
With vendor agnostic best-of-breed and Penguin Computing-designed technology for storage, computing, and networking with either air or liquid cooling
We offer a U.S.-based manufacturing facility with air and liquid cooling and combined total of 6 megawatts of power. More importantly, we have on-site quality assurance engineers who know our technology intimately and test every system personally.
Secure, well-maintained facilities to host your racks in locations across the U.S. that already house some of our AI clients, removing physical footprint, power and cooling related concerns from the equation
What sort of AI system can we help you create?
Tested and supported reference designs include:
- A very high-performance AI platform featuring the NVIDIA DGX-1™ platform
- A high-end HPC system for customers who want to use NVIDIA-based (Tesla® V100) graphics processing unit (GPU)-accelerated computing or those with central processing units (CPUs)/memory needs
- An alternative, high value AI platform based on the AMD Vega GPU based Instinct™ MI25 GPU accelerator for customers who need single precision-oriented HPC
Choose from best-of-breed or proprietary technology, including:
Successful AI design and production requires expertise in all the key components of a balanced AI system, including storage, networking, and computing, to ensure the system actual achieves the goals for which it was intended. And hosting or managing an AI system requires experts who’ve already been there and done that – at scale.