EOS 4a with EB4400: NVCertified Solution
-
The NVIDIA Certified combination of the OSS EOS Server and OSS EB4400 Expansion exemplifies the power and flexibility of OSS’s AI Compute product strategy. Together they provide the power, scalability and flexibility to run the most demanding AI workflows and have undergone the NVIDIA Certification process to demonstrate their effectiveness in running the extensive library of NVIDIA AI optimized software.
Learn more about the EOS Server and the EB4400.
The EOS Server is designed specifically to support large scale-out configurations of connected GPUs in cable attached expansion enclosures. This includes specialized BIOS to support a large number of downstream GPU and IO devices. The OSS 4400 Expansion platform is an ideal product for rugged space constrained applications that need the power of up to 4 NVIDIA A100 GPUs. Together they are a fully Certified AI compute appliance.
The NVIDIA certification process exercises the performance and functionality of the EOS and EB4400 by running a set of software that represents a wide range of real-world applications in data science, visualization and high-performance computing (HPC). It also covers infrastructure performance acceleration, such as network and storage offload, security features such as Trusted Platform Module (TPM), and remote management capabilities. -
Features:
- Compatibility and Peformance: Through certification customers are assured GPU-optimized software from the NVIDIA® NGC™ catalog and commercially available applications will run—and be confident that they’ll perform well.
- Security: These NVIDIA-Certified system secure workflows by protecting data at the platform, network, and application layers. Whether deployed in a data center or at the edge, customers can be assured that they don’t have to compromise on security when running accelerated applications.
- Scalability: Through NVIDIA-Certification this server and expansion combination are tested to validate both multi-GPU and multi-node performance, as well as the performance of multiple workloads running on a single GPU with Multi-Instance GPU (MIG) technology.