AMAX Accelerates AI Supercomputing Build Outs With NVIDIA DGX™ A100
DGX A100 Delivers Double the Performance at Half the Price
FREMONT, Calif., May 20, 2020 (Newswire.com) - AMAX, a leader in integrated supply chain manufacturing and orchestration, announced today that its HPC and AI Solutions Group unit is offering and deploying the new NVIDIA DGX A100 system as a key building block to accelerate the deployment of AI supercomputing infrastructure. Enterprise data centers, research centers, academia and OEM solution providers can deploy DGX clusters into their environment or build a ready-to-deploy rack elevation through AMAX Rack Integration Hubs in as fast as 1-2 weeks.
Unmatched data center scalability and flexibility
More complex demands on scientific modeling, simulation, machine learning, and data volumes highlight the demand for faster HPC systems that can perform at petascale. NVIDIA DGX A100 is the single platform that brings together training, inference, and analytics into a consolidated system with optimized software that is the foundational building block for AI infrastructure. It's based on the latest Ampere GPU and PCIe Gen4 architecture that packs eight A100 Tensor Core graphics processing units (GPUS) into a 6U form factor with 320 GB of GPU memory, six NVswitch interconnect fabrics, and nine NVIDIA Mellanox 200Gb/s InfiniBand X6 HDR connectivity for a total of 5 petaflops in a single node.
DGX A100 provides two times the performance at half the price as compared to the previous generation DGX-2 system. A DGX SuperPOD rack cluster using 20 DGX A100 interconnected nodes in a five rack footprint obtains 48 petaflops of AI computing power. AMAX offers advanced cooling options and matching Gen4-based building blocks for network and storage to unleash the full performance of DGX A100.
“The secret to breaking the barriers of AI supercomputing for scientists and engineers is not just about packing unprecedented compute density with low latency I/O data interconnects and ultra-fast storage into a rack elevation,” said Lee Yee, SVP of Marketing and Corporate Strategy at AMAX. “NVIDIA is taking AI computing to new levels through the power of collaborative partner ecosystems that work. We have a long partner history together, and our rack integration hubs are ready in capacity to build and integrate the next generation of DGX-based solutions into data centers of all sizes.”
AMAX Rack Integration Hubs (ARIH)
ARIH are meticulously designed for quality engineering in L-12 rack-scale server, storage and networking integration projects while simplifying the time-to-uptime process. They provide complete rack-level thermal testing, advanced cooling solutions, power efficiency testing, burn-in, quality and logistics support. Full racks are tested as an integrated unit to ensure operability at deployment. Customized rack designs can be implemented to provide advanced cooling options, optimize power, network cable routing to fully tailor actual datacenter environments. End-to-end field-proven services from engineering design, site surveys, material preparation and assembly mitigate the risks of introducing new infrastructure requirements. ARIH incorporates rack-scale design that includes power management, security and provisioning. AMAX offers a comprehensive line of GPU-integrated solutions, data center building blocks, and OEM manufacturing services for research labs and data centers worldwide.
AMAX is a leading technology solutions company providing integrated supply chain manufacturing and orchestration services. As a Foxconn Technology Group affiliate, AMAX specializes in the manufacturing of server appliances, HPC and storage, AI computing and analytic platforms, and delivers data center solutions at any scale – on any premise. AMAX local hubs across North America, Europe, and Asia enable customers to move mass customization closer to the edge of business consumption. AMAX creates value for clients through three operating groups: OEM Manufacturing Solutions, HPC and AI Solutions, and Smart Manufacturing Solutions.
Categories: Artificial Intelligence and Expert Systems