Cisco UCSC-GPU-A10 | NVIDIA A10 24GB GDDR6 GPU, 150W Passive, PCIe 4.0

Q: What workloads is the UCSC-GPU-A10 optimized for?
A: The NVIDIA A10 is optimized for virtual desktop infrastructure, AI inference, mainstream graphics rendering, virtual workstations, and compute-intensive applications. It excels in multi-tenant data center environments where efficient GPU virtualization is required.

Q: Which Cisco UCS servers support this GPU?
A: The UCSC-GPU-A10 is compatible with Cisco UCS C220 M6, C225 M6, C240 M6, and C245 M6 rack servers. It requires a unique SBIOS ID configured by Cisco Integrated Management Controller or UCS Manager, so Cisco-supplied cards are recommended for proper functionality.

Q: Does this GPU require active cooling or supplemental power?
A: The A10 uses passive cooling and relies on server chassis airflow to maintain safe operating temperatures. It requires an 8-pin PCIe power connector and consumes up to 150W under full load.

Q: Can I partition this GPU for multiple virtual machines?
A: Yes, with NVIDIA vGPU software licenses, the A10 supports GPU virtualization and partitioning, enabling multiple users or workloads to share GPU resources securely. It is compatible with NVIDIA Virtual PC, Virtual Applications, RTX Virtual Workstation, and Virtual Compute Server software.

Q: What is the benefit of the single-slot form factor?
A: The single-slot design allows for higher GPU density in standard rack servers, enabling multiple GPUs to be installed in a single chassis. This is ideal for scaling VDI deployments, AI inference clusters, or virtualized graphics workloads without requiring specialized wide-slot chassis.