Skip to product information
1 of 1

Cisco

Cisco UCSX-GPU-A40-D | NVIDIA A40 48GB GDDR6 GPU, Passive, PCIe Gen4

SKU:UCSX-GPU-A40-D

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 We do not quote Gmail, Outlook or Hotmail addresses, or supply for home use. 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 UCS X-Series NVIDIA A40 is a passively cooled, dual-slot data centre GPU built on the Ampere architecture. It is aimed at virtual workstations and VDI, professional visualisation and rendering, and AI, data science and compute workloads running in the data centre.

The card carries 48GB of GDDR6 memory with ECC and 696 GB/s of memory bandwidth, with 10,752 CUDA cores, 84 second-generation RT cores and 336 third-generation Tensor cores. It connects over PCIe Gen4 x16 at 64 GB/s, and NVLink at 112.5 GB/s allows two cards to be paired for up to 96GB of memory. Maximum power draw is 300W through an 8-pin CPU power connector, and cooling is passive, so it relies on host chassis airflow. This -D variant is the passive A40 option with power cable.

The A40 supports NVIDIA vPC/vApps, RTX Virtual Workstation and Virtual Compute Server vGPU software, which is licensed separately from the hardware. It includes secure and measured boot with a hardware root of trust and is NEBS Level 3 ready.

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

Features

- Second-generation RT cores for real-time ray tracing
- Third-generation Tensor cores with structural sparsity for AI training and inference
- 48GB GDDR6 memory, scalable to 96GB with NVLink
- PCIe Gen4 doubles the bandwidth of Gen3
- Hardware root of trust with secure and measured boot
- Supports NVIDIA RTX Virtual Workstation, vPC/vApps and Virtual Compute Server vGPU software
- Passive, power-efficient dual-slot design for data centre deployment

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 architecture: NVIDIA Ampere
- GPU memory: 48GB GDDR6 with ECC
- Memory bandwidth: 696 GB/s
- Host interface: PCIe Gen4 x16, 64 GB/s
- NVLink: 112.5 GB/s (bidirectional)
- CUDA cores: 10,752
- RT cores (2nd generation): 84
- Tensor cores (3rd generation): 336
- Peak FP32: 37.4 TFLOPS
- Peak FP16 Tensor: 149.7 TFLOPS (299.4 with sparsity)
- Peak TF32 Tensor: 74.8 TFLOPS (149.6 with sparsity)
- Peak INT8 Tensor: 299.3 TOPS (598.6 with sparsity)
- RT core performance: 73.1 TFLOPS
- Form factor: 4.4 in (H) x 10.5 in (L), dual slot
- Display ports: 3x DisplayPort 1.4 (disabled by default for virtualisation)
- Maximum power consumption: 300W
- Power connector: 8-pin CPU
- Cooling: Passive
- Video engines: 1x NVENC, 2x NVDEC (includes AV1 decode)
- Compute APIs: CUDA, DirectCompute, OpenCL, OpenACC
- NEBS: Level 3 ready
- MIG support: No

FAQs

Q: Does the A40 include vGPU licensing?
A: No. The GPU hardware does not include vGPU licensing. NVIDIA vPC/vApps, RTX Virtual Workstation and Virtual Compute Server software are licensed separately, and the NVIDIA Virtual GPU Licensing Guide lists the supported profiles.

Q: What power does the A40 require?
A: It draws up to 300W and uses an 8-pin CPU power connector.

Q: Does the A40 have a fan?
A: No. It is passively cooled and depends on airflow from the host chassis, so it must be installed in a system designed to push air across the card.

Q: Can two A40 cards be linked together?
A: Yes. The A40 supports NVLink at 112.5 GB/s, which allows two cards to be paired for up to 96GB of combined GPU memory.

Q: Does the A40 support MIG?
A: No. Multi-Instance GPU is not supported on the A40. GPU sharing is handled through NVIDIA vGPU software instead.

  • 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