Supercharge your AI and HPC workloads with Supermicro’s latest NVIDIA HGX H100 8-GPU systems. Now available for convenient remote sessions allowing benchmarking of workloads and validation in a test environment through our exclusive JumpStart Program. These systems are engineered to power the next generation of large-scale AI models, inference, and high-performance computing (HPC) workloads.
Designed for scalability and performance, the NVIDIA HGX H100 delivers breakthrough innovations to accelerate your AI operations, featuring cutting-edge technology such as 4th Gen NVLink and PCIe 5.0 connectivity. Discover how the power of 8 NVIDIA HGX H100 GPUs can transform your AI infrastructure today.
Key Features of the Supermicro NVIDIA HGX H100 System:
- 8 NVIDIA HGX H100 GPUs: Each with 80GB SXM5 for a total of 640GB HBM3 memory, delivering exceptional speed for training and inference.
- 4th Gen NVLink: Seamlessly interconnect GPUs with blazing-fast data transfer speeds.
- PCIe 5.0 Connectivity: Offers up to 800Gb/s networking capabilities for unparalleled data throughput.
- 1:1 GPU to NIC Support: Ensures efficiency for large-scale AI and HPC workloads.
- Advanced Cooling and Power Efficiency: Optimised for high performance under heavy workloads, without compromising on energy efficiency.
| CPU | Dual Intel Xeon Platinum 8568Y+ 48-core Processors (96 cores and 192 threads) |
| Drive | 4x KIOXIA CD6 7.68TB NVMe |
| RAM | 32x SK Hynix 64GB DDR5 4800MHz (Total 2TB) |
| GPU | NVIDIA HGX 8-GPU 80GB SXM5 (Total 640GB HBM3) |
Benefits of Using Supermicro NVIDIA HGX H100 Systems:
- Optimised for AI Training and Inference: Delivers the speed and scale needed for large AI models.
- High-Performance Connectivity: Leverages the latest interconnect and networking standards.
- Energy Efficient: Designed to provide high performance with optimised energy consumption.
- Simplified Remote Access: Use the JumpStart Program to validate your workloads straight from your browser.
Apply Now to Secure Your Access
Ready to experience the future of AI infrastructure? Apply for our JumpStart Program to begin testing your AI and HPC models on Supermicro’s NVIDIA HGX H100 8-GPU systems. Convenient remote sessions allowing benchmarking of workloads and validation in a test environment
How it Works
1. Sign Up
First, please sign up and register for the JumpStart program here via Exertis Enterprise. Simply fill out your name, email, and briefly let us know what you plan to do with the system.
2. Get Qualified
The Exertis Enterprise team will process this directly with Supermicro and you will receive a welcome email where you will need to verify your NDA information (Supermicro special NDA and Intel CNDA) in order for us to schedule your access to the JumpStart server.
3. Access
You will provided with a unique username, password, and a URL to enable your access to the JumpStart server. Also, Supermicro would love to receive your feedback on the program and will send you a survey once your access period is completed.
