Cisco UCSC-885A-M8-H23 modular server chassis 3.84 TB

2x AMD EPYC 9575F 3.3 GHz (Max Boost 5 GHz) CPUs, 8x NVIDIA HGX H200 SXM GPUs, 24x 96GB up to 6,000MT/s DIMMs, 2x 960GB M.2 NVMe Boot Drive, 2x 2.5” 1.92TB NVMe SSD Server Drives, 8x NVIDIA ConnectX-7 (1x400G) for East-West N/W, 2x NVIDIA ConnectX-7 (2x200G) for North-South N/W, 1x Intel X710-T2L

The Cisco® UCS C885A M8 Rack Server is a dense GPU server that delivers massive, scalable performance for AI workloads, such as Large Language Model (LLM) training, fine-tuning, large model inferencing, and Retrieval Augmented Generation (RAG).

Built on the NVIDIA HGX platform, the Cisco UCS C885A M8 Rack Server is a dense-GPU server designed to deliver scalable accelerated compute capabilities to address the most demanding AI workloads, including large deep learning/Large Language Model (LLM) training, model fine-tuning, large model inferencing, and Retrieval-Augmented Generation (RAG).

The server offers a choice of 8 NVIDIA HGX H100 or H200 Tensor Core GPUs, or 8 AMD MI300X and MI350X OAM GPUs to deliver massive, accelerated computational performance in a single server, as well as one NVIDIA ConnectX-7 NIC or NVIDIA BlueField-3 SuperNIC per GPU to scale AI model training across a cluster of dense GPU servers.

The server is managed by Cisco Intersight®, which can help reduce your Total Cost of Ownership (TCO) and increase your business agility.

Note: Initially, the local server management interface will handle configuration and management, while Cisco Intersight will provide inventory capabilities through an integrated Device Connector. Full management operations and configurations through Cisco Intersight will be introduced shortly thereafter in a subsequent phase.

The server is offered in fixed configurations that are optimized for intensive AI and HPC workloads.