• Samsung Galaxy S20 Ultra 5G smartphone with Cisco HX-GPU-A10 GPU
1 1

Cisco

Cisco HX-GPU-A10 | NVIDIA A10 GPU, 24GB GDDR6 ECC, 9216 CUDA Cores, 150W

HX-GPU-A10

Stock Status: Enquire

Request Quote
Sale Sold out
Shipping calculated at checkout.

Description

The Cisco HX-GPU-A10 is an NVIDIA A10 Tensor Core GPU designed for enterprise AI inference, virtual workstations, and data center graphics workloads. Built on the Ampere architecture with 9,216 CUDA cores, 72 RT cores, and 24GB of GDDR6 ECC memory, it delivers versatile acceleration for rendering, AI compute, and VDI in a single-slot, passively-cooled form factor. Ideal for Cisco HyperFlex HX240c and HX220c M6 server platforms, the A10 combines mainstream GPU performance with data center efficiency at 150W TDP.

Features

NVIDIA Ampere architecture with 9,216 CUDA cores for versatile compute, graphics, and AI workloads
- Third-generation Tensor Cores with TF32, FP16, and INT8 precision for accelerated AI inference
- 72 second-generation RT cores for real-time ray tracing and advanced rendering
- 24GB GDDR6 ECC memory with 600 GB/s bandwidth for large datasets and models
- PCIe Gen4 x16 interface delivering 64 GB/s bidirectional bandwidth
- Single-slot, full-height, full-length form factor optimized for dense data center deployments
- Passive cooling design for quiet operation in enterprise server chassis
- 150W TDP for energy-efficient performance in mainstream servers
- Hardware-accelerated video encoding and decoding with AV1 decode support
- Secure boot with hardware root-of-trust for firmware integrity
- NVIDIA vGPU software support for multi-tenant VDI and virtual workstation environments
- Compatible with NVIDIA NGC catalog, CUDA-X libraries, and 1,800+ GPU-optimized applications
- Mixed-precision compute capabilities including FP32, TF32, FP16, and INT8/INT4 operations
- Suitable for AI inference, deep learning, graphics virtualization, and compute 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

Technical Specifications

FAQs

Technical Specifications

GPU Architecture: NVIDIA Ampere
- CUDA Cores: 9,216 FP32 cores
- Tensor Cores: Third-generation Tensor Cores
- RT Cores: 72 second-generation RT cores
- GPU Memory: 24GB GDDR6 with ECC
- Memory Bandwidth: 600 GB/s
- Memory Interface: 384-bit
- System Interface: PCIe Gen4 x16
- Form Factor: Single-wide, full-height, full-length (1-slot FHFL)
- Cooling: Passive (requires chassis airflow)
- Power Consumption: 150W TDP
- Power Connector: Single 8-pin PCIe
- Performance: 31.2 TFLOPS (FP32), 62.5 TFLOPS (TF32), 125 TFLOPS (FP16)
- Encoder/Decoder: 1x encoder, 1x decoder with AV1 decode support
- Compatible Platforms: Cisco HyperFlex HX240c M6, HX220c M6, HX245c M6 server nodes
- vGPU Software Support: NVIDIA vPC, vApps, RTX vWS, Virtual Compute Server

FAQs

Q: Which Cisco HyperFlex platforms support the HX-GPU-A10?
A: The HX-GPU-A10 is supported on Cisco HyperFlex M6 platforms including HX240c M6, HX220c M6, and HX245c M6 nodes. It can be installed in PCIe Gen4 x16 or x8 slots across multiple risers, with some configurations supporting up to five A10 GPUs per node. Compatible with HX Data Platform Release 5.0(2b) and later.

Q: What are the primary use cases for the NVIDIA A10 GPU?
A: The A10 is designed for AI inference workloads, virtual desktop infrastructure (VDI), graphics-rich virtual workstations running CAD and design applications, rendering farms, data science, and video encoding/decoding. It bridges the gap between dedicated inference accelerators and high-end training GPUs, offering versatile performance for multi-tenant data center environments.

Q: Does the HX-GPU-A10 require special power or cooling?
A: The A10 uses passive cooling and requires a single 8-pin PCIe power connector. It must be installed in a server chassis with adequate airflow to dissipate the 150W thermal load. Cisco HyperFlex servers with compatible riser configurations and proper heatsinks (low-profile heatsink HX-HSLP-M6) provide the necessary cooling infrastructure.

Q: Can the A10 GPU be used for virtualized workloads?
A: Yes, the A10 supports NVIDIA vGPU software including vPC, vApps, RTX Virtual Workstation (vWS), and Virtual Compute Server. This enables GPU sharing across multiple virtual machines for VDI, virtual workstations, and compute workloads, making it ideal for multi-tenant enterprise and cloud environments.

Q: How does the A10 compare to other NVIDIA data center GPUs?
A: The A10 offers a balance between the entry-level T4 and high-end A100. It provides over 2.5x more Tensor performance than the T4, supports PCIe Gen4 for increased bandwidth, and includes 24GB of memory suitable for medium-sized AI models and graphics workloads. Its 150W power envelope makes it more efficient than higher-performance GPUs while maintaining strong inference and visualization capabilities.

Recently Viewed

  • 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