Penguin Solutions, an SGH brand, announced it has been selected by Voltage Park to provide core professional and managed services and software for its large-scale NVIDIA-based clusters. Penguin designs, builds, deploys, and manages AI and accelerated computing infrastructures at scale. Voltage Park is a next-generation cloud company focused on providing accessible machine learning (ML) infrastructure for AI.
Voltage Park’s cloud environment is one of the more significant ML compute infrastructures in the world. Under the architecture being deployed, each compute node instance features eight NVIDIA H100 Tensor Core GPUs for a total of 640GB of GPU memory per node. A high-performance, low-latency fabric built with NVIDIA InfiniBand Networking ensures workloads can scale across clusters of interconnected systems, allowing multiple instances to act as one massive GPU to meet the performance requirements of advanced AI training. High-performance storage is also being integrated to provide a complete solution for AI supercomputing.
“Voltage Park was in search of help to manage and maximize the efficiency of their large AI infrastructure,” said Pete Manca, President of Penguin Solutions. “As a multi-tenancy cloud provider, they offer customers several choices for computing power consumption. They turned to Penguin as an expert partner with large scale infrastructure experience to ensure maximum GPU performance and cluster availability that meets the needs of their compute-hungry customers.”
Penguin, a certified NVIDIA DGX-ready Managed Services Provider, will integrate and ensure production readiness for the majority of Voltage Park’s 24,000 GPU deployment of NVIDIA H100s and next-generation 3.2 TB InfiniBand and Ethernet interconnects. The deployment process will follow Penguin’s OriginAI® solution, a proven ‘Design. Build. Deploy. Manage.’ methodology with an objective of achieving a rapid and high-quality result. Penguin is also providing professional services and managed services, and the entire platform which spans four data centers will be powered by Penguin’s Scyld ClusterWare® software platform.
“Penguin’s track record of successfully deploying and managing large AI factories was compelling, but it was their Scyld ClusterWare software coupled with their services offerings that were truly pivotal to our decision,” said Ozan Kaya, CEO Voltage Park. “We are excited to bring customers AI infrastructure through our unique combination of flexible, exchange-based pricing for short-term compute access alongside traditional long-term rentals. Penguin’s end-to-end ability to deliver, optimize, and support the complete environment for multi-tenancy is helping bring our vision to life.”
SOURCE: BusinessWire