PORTFOLIO EXPANSION TO INCLUDE FULLY INTEGRATED WITH NVIDIA A100 GPU-POWERED SYSTEMS

For the most demanding AI workloads, Supermicro builds the most powerful and fastest servers based on NVIDIA A100™ Tensor Core GPUs. Supermicro supports a range of customer requirements with optimised systems for the new HGX™ A100 8-GPU and HGX™ A100 4-GPU platforms. With the latest version of NVIDIA® NVLink™ and NVIDIA NVSwitch™ technologies, these servers can deliver up to 5 PetaFLOPS of AI performance in a single 4U system. Supermicro can also support the new NVIDIA A100 PCI-E GPUs in a range of systems, with up to 8 GPUs in a 4U server.

GET TO KNOW THE BENEFITS

The latest high-density 2U and 4U servers leverage Supermicro's advanced thermal design, including custom heatsinks and optional liquid cooling, and feature NVIDIA HGX A100 4-GPU 8-GPU baseboards and a new 4U server that supports eight NVIDIA A100 PCI-E GPUs . Supermicro's Advanced I/O Module (AIOM) form factor further enhances network communications with high flexibility. The AIOM can be coupled with the latest high-speed PCI-E 4.0 low-latency storage and networking devices supporting NVIDIA GPUDirect ®RDMA and GPUDirect Storage with NVME over Fabrics (NVMe-oF) on NVIDIA Mellanox® InfiniBand, which provides the expandable multi-GPU system with continuous data flow without bottlenecks. In addition, Supermicro's Titanium Level power supplies with the industry's highest efficiency of 96% provide even more cost savings while enabling redundant support for the GPUs.
"Supermicro systems powered by the NVIDIA A100 can quickly scale to thousands of GPUs. With the new multi-instance GPU technology, each A100 GPU can be partitioned into seven isolated GPU instances to run different jobs," said Paresh Kharya, senior director of product management and marketing at NVIDIA. "NVIDIA A100 Tensor Core GPUs with TensorFloat 32 provide up to 20 times the compute capacity of the previous generation without requiring code changes."
Supermicro continues its support of NVIDIA's advanced GPUs in a variety of form factors optimised for customers' specific application scenarios and requirements. The 1U GPU systems include up to four NVIDIA GPUs with NVLink, including the NEBS Level 3 certified, 5G/Edge-ready SYS- 1029GQ. Supermicro's 2U GPU systems, such as the SYS- 2029GP-TR, can support up to six NVIDIA V100 GPUs with dual PCI-E root complex capability in one system. Finally, the 10U GPU servers, such as the SYS- 9029GP-TNVRT, support 16 V100 SXM3 GPU expansions with up to two Intel Xeon Scalable processors with integrated AI acceleration.

The flexible range of solutions with NVIDIA GPUs and GPU software from the NVIDIA NGC™ ecosystem provides the right building blocks for different tasks for companies in different industries - from AI inferencing on evolved models to HPC to high-end training

DENSE, VERSATILE 4U DESIGN WITH HGX A100 8-GPU

The new 4U GPU system features the NVIDIA HGX A100 8-GPU baseboard, up to six NVMe U.2 and two NVMe M.2, ten PCI-E 4.0 x16 slots. Supermicro's unique AIOM support enlivens the 8-GPU communication and data flow between systems through the latest technology stacks such as GPUDirect RDMA, GPUDirect Storage and NVMe-oF on InfiniBand. The system uses NVIDIA NVLink and NVSwitch technology. It is ideal for large-scale deep learning training, neural network modelling applications for research or national labs, supercomputing clusters and HPC cloud services.

Supermicro SuperMinute VIDEO:
​​​​​​​4U System with HGX A100 8-GPU

Ideal for large-scale deep learning training and neural network modelling applications. The new 4U GPU system features the NVIDIA HGX A100 8-GPU baseboard, up to six NVMe U.2 and two NVMe M.2, 10 PCI-E 4.0 x16 I/O, with Supermicro's unique AIOM support strengthening 8-GPU communication and data flow between systems through the latest technology stacks such as NVIDIA NVLink and NVSwitch, GPUDirect RDMA, GPUDirect Storage and NVMe-oF on InfiniBand.

HYPER-TIGHT, ENERGY-EFFICIENT 8U SUPERBLADE WITH 20 A100 PCI-E GPUS

Super Micro Computer doubles down on GPU capabilities with a new 4U server that supports eight NVIDIA® HGX A100™ GPUs. With the new NVIDIA HGX A100™ 8-GPU server, Supermicro now offers the industry's broadest and deepest selection of GPU systems to run applications from the edge to the cloud.

The industry's highest density GPU blade server can support up to 20 nodes and 40 GPUs with two single-width GPUs per node or one NVIDIA Tensor Core A100 PCI-E GPU per node in Supermicro's 8U SuperBlade enclosure. The 20 NVIDIA A100 GPUs in 8U increase the density of computing power in a smaller footprint and provide customers with total cost of ownership savings. To support the GPU-optimised configuration and maintain the peak performance and throughput required for demanding AI applications, the SuperBlade provides 100% non-blocking HDR 200Gb/s InfiniBand network infrastructure to accelerate Deep Learning and enable real-time analytics and decision making. High density, reliability and upgradeability make the SuperBlade a perfect building block for enterprise applications to deliver AI-powered services.

SUPERMICROS ADVANCED 2U DESIGN FOR HGX A100 4-GPU

This 2U system features the NVIDIA HGX A100 4-GPU baseboard with Supermicro's advanced heat sink design to maintain optimal system temperature under full load, all in a compact form factor. The system enables high GPU peer-to-peer communication via NVIDIA NVLink®, up to 8TB of DDR4 3200Mhz system memory, five PCI-E 4.0 I/O slots supporting GPUDirect RDMA, and four hot-swappable NVMe with GPUDirect storage capability.

supermicro SuperMinute VIDEO:
​​​​​​​2U System with HGX A100 4-GPU

This 2U system features the NVIDIA HGX A100 4-GPU baseboard with Supermicro's advanced heat sink design to maintain optimal system temperature under full load, all in a compact form factor. The system enables high GPU peer-to-peer communication via NVIDIA NVLink®, up to 8TB of DDR4 3200Mhz system memory, five PCI-E 4.0 I/O slots supporting GPUDirect RDMA, and four hot-swappable NVMe with GPUDirect storage capability.
"These new Supermicro systems boost overall performance for accelerated workloads needed for fast-changing
​​​​​​​markets, including HPC, data analytics, deep learning training and inference."