8-GPU server with NVIDIA H100 with NVLink GPUs – HGX H100™ 8 GPU
Microway’s Octoputer™ 8 GPU server with NVIDIA NVLink® Technology provides eight NVIDIA H100 Tensor Core GPUs and two 4th Gen Xeon Scalable Processors within a 6U rackmount server platform. This system is based on the NVIDIA HGX H100 8 GPU platform.
Leveraging NVIDIA’s Fourth Generation NVIDIA NVLink GPU interconnect, this system allows applications to utilize the memory of all eight GPUs combined in a single memory space. The architecture is uniquely suited to AI Training and highly GPU accelerated HPC applications.
Each Octoputer server arrives fully integrated with the operating system of your choice, along with the NVIDIA CUDA toolkit. Your choice of libraries, tools, and software applications including NVIDIA AI Enterprise are available for an additional fee. All systems pass rigorous stress tests for over 72 hours prior to shipment.
Need scalable compute? Microway provides turn-key GPU clusters including with InfiniBand interconnects and GPU-Direct RDMA capability.
Features
- 8 NVIDIA H100 GPUs with: 80GB HBM3 memory, 4th Gen NVIDIA NVLink Technology, and 4th Gen Tensor Cores with a new transformer engine
- 4x 3rd Gen NVIDIA NVSwitches for maximum GPU-GPU Bandwidth (7.2TB/sec of total BW) – Full all-to-all communication with 900GB/sec of bandwidth per GPU
- Supports GPUDirect® RDMA over PCI-E x16 5.0 to NVIDIA ConnectX-7 NDR 400Gb InfiniBand adapters
- Up to 120 cores from 2x 4th Gen Xeon Scalable Processors
- Up to 8 TB system memory
- Additional slots for NDR InfiniBand, 25/50/100G Ethernet, and high-speed storage
- NVIDIA Certified System Configuration for the ultimate in performance and validation. NVIDIA AI Enterprise-ready.
- NVIDIA CUDA Toolkit installed and configured – ready to run GPU jobs!
Specifications
- (2) Intel Xeon 4th Gen Xeon Scalable Processor “Sapphire Rapids” CPUs (clock speed: up to 3.7 GHz)
- Eight NVIDIA H100 GPUs via SXM5 GPU modules, each with 450GB/s of NVLink interconnect bandwidth (900GB/s bidirectional)
- Eight-Channel DDR5 4800 MHz ECC/Registered Memory (32 slots)
- Eight PCI-Express x16 5.0 low-profile, half-length slots for GPUDirect RDMA enabled InfiniBand
- Up to (24) Hot-Swap 2.5” NVMe drives
- Two PCI-Express x16 5.0 low-profile, half-length slots for storage or additional fabric
- IPMI 2.0 with Dedicated LAN Support
- 12,000W High-Efficiency Power with 3+1 Redundancy
Accessories/Options
- ConnectX-7 400Gb HDR or ConnectX-6 200Gb HDR InfiniBand or 25G/50G/100G Ethernet
- TPM 2.0, with optional TXT support
- NVIDIA HPC SDK for GPUs
- NVIDIA AI Enterprise software suite
- Intel compilers, libraries and tools
Support
Supported for Life
Our technicians and sales staff consistently ensure that your entire experience with Microway is handled promptly, creatively, and professionally.
Telephone support is available for the lifetime of your server(s) by Microway’s experienced technicians. After the initial warranty period, hardware warranties are offered on an annual basis. Out-of-warranty repairs are available on a time & materials basis.
Price
Contact us for pricing
Each Microway system is customized to your requirements. Final pricing depends upon configuration and any applicable educational or government discounts.
