Skip to product information
1 of 1

Cisco

Cisco UCSC-885A-M8-H12 | UCS C885A M8 AI Server, 8x H100 GPU, 2.3TB RAM

SKU:UCSC-885A-M8-H12

Request Quote
Sale Sold out
Shipping calculated at checkout.

Stock Status: Enquire

 Trade supply only. Server & GPU quote requests need a company name, ABN and work email address. We do not quote Gmail, Outlook or Hotmail addresses, or supply for home use. 

20-min response

Need a quote or bulk pricing? Answered within 20 minutes, Sydney business hours.

Custom Built Servers

Need a different spec, quantity or configuration?

We build configured-to-order refurbished servers from Cisco, Dell, HPE, NVIDIA Lenovo and Supermicro. Tell us what you need and we'll come back with options within 24 hours.

Configure your server now →

Description

The Cisco UCS C885A M8 is a dense-GPU rack server from the UCS C-Series, built for demanding AI workloads such as LLM training, model fine-tuning, large-model inferencing, Retrieval-Augmented Generation (RAG) and HPC in dense-GPU clusters.

This SKU is the fixed H100 configuration. It combines eight NVIDIA H100 Tensor Core SXM GPUs on a UBB GPU sled with two AMD EPYC 9575F CPUs and 24 x 96GB DDR5-6400 DIMMs, giving 2.3TB of memory. Eight NVIDIA BlueField-3 B3140H SuperNICs carry east-west GPU traffic, while one BlueField-3 B3220 handles north-south traffic. Storage is 16 x 1.92TB NVMe SSDs, plus an Intel X710 T2L 10GbE OCP NIC. Power comes from six 3000W 54V AC supplies for the GPU tray and two 2700W 12V AC/DC Titanium supplies for the CPU tray, in a redundant design, with 12 fans for cooling.

The C885A M8 chassis also supports NVIDIA H200 and AMD Instinct MI300X and MI350X GPU sleds in other configurations. Check NVIDIA's site for H100 availability, as the GPU has reached end of life.

Enquire with XS Network Tech for new or used pricing and availability.

Features

- Eight NVIDIA H100 SXM GPUs on a UBB GPU sled for LLM training, fine-tuning and inferencing
- Eight BlueField-3 B3140H SuperNICs dedicated to east-west GPU traffic
- BlueField-3 B3220 for north-south traffic
- OCP 2.0-based DC-SCM module as the default primary management interface with a common security model
- Three groups of hot-swappable NVMe SSD storage
- Redundant power supply design across GPU and CPU trays
- Fixed, pre-optimised configuration for AI and HPC workloads

Warranty

All products sold by XS Network Tech include a 12-month warranty on both new and used items. Our in-house technical team thoroughly tests used hardware prior to sale to ensure enterprise-grade reliability.

All technical data should be verified on the manufacturer data sheets.

View full details

specs-tabs

Collapsible content

Technical Specifications

FAQs

Technical Specifications

- PID: UCSC-885A-M8-H12
- GPU: 8 x NVIDIA H100 Tensor Core SXM GPUs (C885A-M8-H1SX-SLD UBB GPU sled)
- CPU: 2 x AMD EPYC 9575F (CPU sled C885A-M8-CC-SLD02)
- Memory: 24 x 96GB DDR5-6400 DIMMs (2.3TB)
- Storage: 16 x 1.92TB NVMe SSDs (C885A-NV-C1T9KV)
- East-west NICs: 8 x NVIDIA BlueField-3 B3140H SuperNICs
- North-south NIC: 1 x NVIDIA BlueField-3 B3220
- OCP NIC: 1 x Intel X710 T2L 10GbE (C885A-O-ID10GC)
- GPU tray power: 6 x 3000W 54V AC power supplies
- CPU tray power: 2 x 2700W 12V AC/DC Titanium power supplies
- Management module: DC-SCM card, AMD (C885A-M8-DCSCM)
- Management ports: 1 x 1GbE RJ45, USB 3.2 Gen1x1 Type-C and Type-A, USB Type-C debug, Mini DisplayPort
- Front panel: power button with LED, UID button, 2 x USB 2.0, thermal sensor
- Cooling: 12 fans
- Chassis: C885A-M8-CHASSIS with C885A-HCI-BZL bezel
- Optional boot device: Cisco M.2 NVMe BOOT RAID Controller (UCS-M2-NVRAID), PCIe Gen4 x8, 2 x 960GB M.2 NVMe SSDs, RAID 0/1

FAQs

Q: What workloads is this server designed for?
A: It targets deep learning and LLM training, model fine-tuning, large-model inferencing, Retrieval-Augmented Generation (RAG) and HPC in dense-GPU AI clusters.

Q: How is power supplied to the server?
A: Six 3000W 54V AC power supplies feed the GPU tray and two 2700W 12V AC/DC Titanium power supplies feed the CPU tray, in a redundant design.

Q: How does this SKU differ from other C885A M8 configurations?
A: This SKU is the H100 configuration with 8x B3140H SuperNICs, 1x B3220 and 2.3TB of memory. The same chassis supports H200 and AMD Instinct MI300X or MI350X GPU sleds in other configurations.

Q: How is network traffic handled?
A: Eight BlueField-3 B3140H SuperNICs carry east-west GPU traffic and one BlueField-3 B3220 carries north-south traffic. An Intel X710 T2L 10GbE OCP NIC is also fitted.

Q: Is the H100 GPU still available from NVIDIA?
A: Cisco notes NVIDIA H100 GPUs have reached end of life, so check NVIDIA's site for current availability.

  • Request a Quote

    Looking for competitive pricing? Submit a request, and our team will provide a tailored quote that fits your needs.

  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Call us Now  
  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Contact us