Penguin Solutions today announced the expansion of its OriginAI offering to include validated, pre-defined AI architectures incorporating NVIDIA technology. It includes Penguin’s cluster management software and managed services.
OriginAI infrastructure is intended to streamline AI implementation and management to enable predictable AI performance from clusters ranging from hundreds to thousands of GPUs.
Penguin said OriginAI provides “assured infrastructure for critical, demanding workloads by combining proven architectures, latest generation hardware, advanced cluster management software, and expert professional services.”
These architectures are based on one-pod, four-pod, and 16-pod configurations that can scale from 256 to more than 16,000 GPUs. OriginAI incorporates NVIDIA H100 GPUs, Penguin’s Scyld ClusterWare 12.2 software, and networking and storage options.
“Better GPU performance and controlled costs are top of mind for customers today,” said Matt Eastwood, SVP of enterprise infrastructure research at IDC. “Validated, scalable OriginAI architectures developed with Penguin Solutions’ hands-on expertise in designing, integrating, installing and provisioning AI infrastructure deliver both. Penguin continues to enable AI workloads for the most demanding environments through its innovative high-performance solutions and services.”
OriginAI utilizes Penguin’s in-factory burn-in and integration environment to validate AI cluster performance prior to shipment. This combination “delivers greater than 95 percent overall cluster efficiency while driving higher GPU throughput than traditional approaches,” according to Penguin.
“Designing, deploying, and operating AI factories is an incredibly complex endeavor. Our OriginAI solution builds on Penguin’s extensive AI infrastructure expertise to reduce this complexity and accelerate return on investment,” said Penguin Solutions President Pete Manca. “Our OriginAI solution is a major step forward in providing CEOs and CIOs the essential and reliable infrastructure they need to deploy and manage demanding AI workloads at scale.”
Penguin Solutions is an NVIDIA-certified Elite OEM and DGX AI Compute Systems Solution Provider and DGX-Ready Managed Services partner that has been delivering AI factories at scale since 2017. The company said that with 25-plus years of HPC experience – and more than 75,000 GPUs deployed and managed to date – Penguin is a partner for AI and HPC solutions and services for such organizations as Georgia Tech, Meta, Sandia National Laboratories and the U.S. Navy.
