June 19, 2024
news
FREMONT, Calif. – Penguin Solutions™, an SGH™ brand, announced it is expanding its OriginAI® solutions to include validated, pre-defined AI architectures incorporating NVIDIA technology.
Backed by Penguin's intelligent cluster management software and expert services, the OriginAI infrastructure is designed to streamline AI implementation and management, delivering predictable AI cluster performance that supports customer ROI from clusters that scale from hundreds to thousands of GPUs.
According to the company, OriginAI solutions combine proven architectures, latest generation hardware, advanced cluster management software and professional services to deliver a proven infrastructure for critical and demanding workloads. These architectures are based on 1-, 4- and 16-pod configurations and can scale from 256 to over 16,000 GPUs. OriginAI solutions incorporate NVIDIA H100 GPUs, Penguin's Scyld ClusterWare® 12.2 software, networking and storage options, and are supported by Penguin's comprehensive managed services.
OriginAI solutions utilize Penguin's in-factory burn-in and integration environment to validate AI cluster performance and ensure production readiness before shipping. By combining these architectures with Penguin's Scyld ClusterWare software and professionally managed services that predictively monitor cluster health and manage the solution's throughput, OriginAI achieves higher GPU throughput than traditional approaches while delivering overall cluster efficiency of over 95%, according to the company.
Penguin Solutions is an NVIDIA-certified Elite OEM, DGX AI Computing System Solution Provider, and DGX-Ready Managed Services Partner.
For more information, visit https://www.penguinsolutions.com/solutions/ai/originai.
