While this isn’t purely a SmartNIC-related article, it is something worthy of sharing that is #SmartNIC adjacent. Last month, after listening to Jensen from Nvidia talking about OpenClaw in his keynote, I thought it was time to begin seriously exploring AI Inference using a GPU installed in a dedicated server at home. Yes, I’m that kind of nerd, with a 22U rack in a closet in my home office, that has a dedicated 25A service, and a hard-wired GbE connection via Google Fiber. Now you’re going to laugh, this new project had to be done on a tight budget, so I repurposed an AMD EPYC Rome 16-core server with 64 GB of RAM that I already had and picked up an AMD Radeon RX 7600 with 8 GB of VRAM for $245 on eBay. As shown in the rightmost yellow column above.
LATEST PODCASTS
07.29.26 Optimizing IDS Performance Using FPGA SmartNICs
05.07.26 Ultra-low Latency Inference at the Network Edge with Xelera
04.16.26 Alex Stein of Liquid Market Solutions Talks about Network Attached Compute
Recent posts
Quote of the month
“Agentic AI turns enterprise data into a living, real-time system — and that system must be protected where data moves, where context is stored and where agents act. With Vera BlueField-4 (BF-4) STX, Nvidia and its ecosystem are building secure-by-design storage infrastructure that enforces trust in silicon at the speed of AI.‘’
— Jensen Huang, CEO/Founder NVIDIA
About
A site focused on the latest advances in SmartNIC and DPU technology.
Leave a Reply