Supermicro is dedicated to bringing market-first innovations in enterprise, cloud, artificial intelligence, and enterprise 5G telco architecture. They supply cutting-edge, high-volume motherboard, energy, and chassis technologies; they are evolving into an HPC solution firm with eco-friendly enterprise servers, HPC clusters, memory, IoT, and switching solutions. Supermicro provides a variety of services such as:
Cloud computing and data centers include:
• The data center server and the virtualized 5G Core are supported by disaggregated servers built on Resource-Saving Infrastructure.
• High-density Blade devices with multi-node Twin design for lower TCO and PUE
• Extensive experience with virtualization and container technologies
• GPU servers with the highest performance for data center server learning
Edge and virtualized RAN, which comprises of:
• FPGA accelerators for 5G RAN, including O-RAN, may be built for greater-performance servers that support up to 3rd Gen Intel Xeon Scalable CPUs.
• Edge AI inference with GPU card possibilities improves local decision-making for time-sensitive applications.
• Telecom-standard components for central offices and tiny data centers, such as NEBS Level 3 compliance, AC/DC power, front I/O, and short-depth servers, as well as IP65 ruggedized chassis for outdoor deployments
IoT and customer premises like:
• Small Edge servers and gateways for Cloud-based IoT device deployment management
• Options for networking that are quickly reconfigurable to enable almost any installation
• A broad selection of VNFs supported by software partners for SD-WAN and uCPE
New NVIDIA H100 units, which include:
8-User Next-Generation Universal GPU System, More I/O, and Vast Storage. It has the following features:
• NVIDIA HGX H100 8-GPU as the GPU
• GPU feature set: The H100 is the most powerful processor ever made and provides up to nine times quicker performance for AI training thanks to its 80 billion transistors.
• Dual processors: CPU
• ECC DDR5 memory with a 4800MT/s speed
• Up to 24 Hot-Swap NVMe/SATA drives
Next-Gen 4U/5U Integrated GPU System
Designed to be modular for maximum versatility. It has the following features:
• NVIDIA HGX H100 4-GPU as the GPU
• GPU Features: Compared to prior generations, the H100 HGX can speed up AI Inference by up to 30 times.
• Dual processors: CPU
• ECC DDR5 memory with a 4800MT/s speed
• Drives: 10 Hot-Swap 2.5″ SATA/SAS or up to 8 Hot-Swap NVMe U.2 connect to PCI-E Switch
Next-Generation 4U 10GPU Gen 5 System
Its flexible design Supports Up to 10 NVIDIA GPUs for AI and Graphically-Intensive Workloads. It has the following features:
• GPU: Up to 10 double-width PCI-E GPUs per node
• GPU feature set: This System’s NVIDIA L40 PCI-E GPUs are excellent for accelerating media and graphic applications.
• Dual processors: CPU
• ECC DDR5 memory with a 4800MT/s speed
• 24 Hot-Swap Bays for Drives
Next-generation 4U 4GPU System
It is geared for data analysts, content producers, and collaboration in the 3D Metaverse. Both dual processor rackmount and Workstation design factors are available. It has the following features:
• NVIDIA PCI-E H100 4-GPU as the GPU
• NVIDIA H100 GPUs boost trust in security cooperation.
• Dual processors: CPU
• ECC DDR5 memory with a 4800MT/s speed
• Drives: Up to 8 NVMe drives, 8 Hot-Swap 3.5″ disk slots, and 2x M.2
The second-generation NVIDIA OVX Reference Model Systems
It is a newly developed next-generation system geared toward driving photorealistic, immersive 3D models, simulators, and digital twins. It has the following features:
• 8 NVIDIA L40 GPUs
• GPU Features: The fastest 3rd generation ray tracing graphics (RT Cores) and 4th generation Tensor Cores are found in the NVIDIA Ada Lovelace architecture.
• Dual processor: CPU
• ECC DDR5 memory with a 4800MT/s speed
• 3 ConnectX-7 SmartNICs for 100G/200G/400G networking on the network
Supermicro can customize HPC systems to match many workloads, requiring high throughput GPUs, compute-intensive applications, or high-capacity storage applications utilized in many sectors.
