Skip to product information
1 of 1

Cisco

Cisco UCSC-GPU-H200-NVL | H200 NVL 141GB HBM3e GPU, 600W, 2-Slot FHFL, C-Series

SKU:UCSC-GPU-H200-NVL

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 We do not quote Gmail, Outlook or Hotmail addresses, or supply for home use. 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 NVIDIA H200 NVL is a double-wide, full-height full-length PCIe GPU accelerator built on the NVIDIA Hopper architecture and supplied as a Cisco UCS C-Series part. It is aimed at enterprise data centres running large language model inference, generative AI training and fine-tuning, and high-performance computing on UCS C-Series rack servers.

The card carries 141GB of HBM3e memory and connects over a PCIe Gen5 x16 host interface. Maximum thermal design power is up to 600W and is configurable. It occupies two slots and supports NVLink bridges for 2-way and 4-way NVL configurations. This is the C-Series (UCSC) version of the part, distinct from the UCSX and HCIX variants used in X-Series and HyperFlex X-Series platforms.

In the Cisco UCS C240 M8 the card is supported in Riser 3C, PCIe slot 7 only, and is power-limited to 450W. A maximum ambient temperature of 28 C applies with the card installed, 256GB DIMMs are not supported alongside it, and CPUs with a TDP above 330W are not supported when a GPU above 75W is fitted. NVIDIA AI Enterprise software is typically licensed separately.

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

Features

- 141GB of HBM3e memory for large language model inference and HPC workloads
- Hopper architecture Tensor Cores with Transformer Engine for generative AI
- NVLink bridge connectivity between GPUs in NVL configurations
- Passively cooled, server-optimised double-wide PCIe card
- PCIe Gen5 x16 host interface for high-bandwidth data transfer
- Suited to generative AI training and fine-tuning in enterprise data centres

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

- GPU: NVIDIA H200 NVL (Hopper) Tensor Core GPU
- Memory: 141GB HBM3e
- Maximum TDP: up to 600W (configurable)
- Host interface: PCIe Gen5 x16
- Form factor: double-wide (2-slot), full-height full-length (FHFL)
- NVLink: bridge support for 2-way and 4-way NVL
- Cisco UCS C240 M8 power limit: 450W with H200 NVL installed
- Cisco UCS C240 M8 slot support: Riser 3C, PCIe slot 7 only
- Cisco UCS C240 M8 ambient temperature limit: maximum 28 C
- Cisco UCS C240 M8 memory restriction: 256GB DIMMs not supported
- Cisco UCS C240 M8 CPU restriction: CPUs above 330W TDP not supported with a GPU over 75W
- Cisco UCS C240 M8 power cable: CBL-G5GPU-C240M7

FAQs

Q: Does this GPU include a software licence?
A: No licence is bundled with this Cisco part number. NVIDIA AI Enterprise software is typically licensed separately.

Q: How much power does it draw?
A: The card has a maximum TDP of up to 600W. In the Cisco UCS C240 M8 it is power-limited to 450W.

Q: Which Cisco server supports this GPU?
A: It is supported in the Cisco UCS C240 M8, in Riser 3C, PCIe slot 7 only, which takes a single full-height full-length double-wide GPU.

Q: Are there any restrictions when it is installed in a UCS C240 M8?
A: 256GB DIMMs are not supported with this GPU. CPUs with a TDP above 330W are not supported when a GPU above 75W is fitted, and the maximum ambient temperature is 28 C.

Q: How does this differ from UCSX-GPU-H200-NVL= and HCIX-GPU-H200-NVL=?
A: UCSC-GPU-H200-NVL is the C-Series version for rack servers. UCSX-GPU-H200-NVL= is used in the UCS X580p PCIe Node for UCS X-Series, and HCIX-GPU-H200-NVL= is used in HyperFlex X-Series.

  • 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