Cisco UCSC-GPU-FLEX170 | Intel Flex 170 GPU, 16GB GDDR6, PCIe 4.0 x16, 150W

Q: What workloads is the Intel Flex 170 GPU optimized for?
A: The Flex 170 is designed for media streaming and transcoding, virtual desktop infrastructure (VDI), AI-powered visual inference, cloud gaming, and content delivery networks. It excels at high-density media processing with hardware-accelerated AV1, H.265, and H.264 codecs.

Q: Does the UCSC-GPU-FLEX170 support GPU virtualization?
A: Yes, the Flex 170 supports SR-IOV (Single Root I/O Virtualization) enabling up to 62 virtual functions. This allows the GPU to be shared across multiple virtual machines for VDI deployments on platforms like VMware ESXi, Linux KVM, Citrix DaaS, and VMware Horizon without per-seat licensing fees.

Q: What Cisco UCS servers are compatible with this GPU?
A: The UCSC-GPU-FLEX170 is compatible with Cisco UCS C-Series rack servers including the C240 M7 (SFF and LFF models) and C240 M8. It requires specific PCIe riser configurations and auxiliary power cables (included with GPU-ready server configurations or available separately).

Q: How does AV1 codec support benefit media workloads?
A: Hardware-accelerated AV1 encoding reduces bandwidth requirements by up to 30% compared to H.264 while maintaining superior video quality. This enables higher streaming density, lower operational costs, and better performance for content delivery, especially valuable for 4K and 8K workflows.

Q: Can the Flex 170 be used for AI inference workloads?
A: Yes, the Flex 170 includes 512 Intel XMX engines optimized for AI matrix operations, delivering up to 500 TOPS for adaptive AI workloads. It's particularly effective for visual AI inference applications such as object detection, image recognition, and real-time video analytics.