Nvidia had its annual GTC gathering of AI developers, fans, media, and gamers recently. The newly announced Nvidia Blackwell platform is scheduled for release in late 2024/early 2025. Blackwell will likely lead to another round of “world domination” – I say that tongue in cheek – for the Nvidia. One thing that struck me is…
The recent popularity and interest in ChatGPT underscore the progress AI has made in affecting everyday life. Who has asked ChatGPT to finish a homework assignment or write a blog? (For full disclosure, this blog has been written by a real person. Or has it? 😉) In any case, the backbone of chatbots like ChatGPT…
Something occurred to me recently. AI hardware system architectures are really applications-specific. Computing platforms from laptops, desktops, workstations and servers follow classical architectures. Inputs, outputs, a central processing unit, busses (control, data and memory), data memory and program or instruction memory all balance in harmony in so many facets of our daily lives. Granted, accelerators…
This video, from the AI Hardware Summit, shows a prototypical Gen-Z PECFF demonstration based on PCIe 6.0, at 64 GT/s. It shows PCB connectors and twinax cable systems, with performance capabilities eminently suitable for AI hardware applications. It’s a Gen-Z backplane with Gen-Z Add-In-Cards (AIC) with top card connectivity. This emulates industry standard AI Acceleration…
AI continues to shake up the tech industry. Nvidia’s plans to buy UK-based Arm Holdings from Japan-based SoftBank illustrates that point. Nvidia GPU chipsets make-up the majority of data-center focused AI hardware applications. The Arm Holding licensee ecosystem offers a new territory for Nvidia to harness the power of AI. Just about every mobile device…
2001: A Space Odyssey. The Terminator. The Matrix. Avengers: Age of Ultron. These and many other Hollywood blockbusters foreshadow the worst fears of some: humankind being cast aside into the ash heap of history by AI-based machine overlords. Is that reality? Not likely, but some AI researchers feel it COULD happen in the decades or…
The FPGA (or ACAP) universe gathered at the San Jose Fairmount last week during the Xilinx Developer Forum. Engineers, data scientists, analysts, distributors, alliance partners and more came to learn about the latest hardware, software and system level solutions from Xilinx. Xilinx CEO Victor Peng set the theme of the event with his keynote. Xilinx…
“Artificial Intelligence” and “AI” are all the rage. From high frequency trading on Wall Street to improved drug development, AI affects more aspects of human lives daily. Many AI solutions harness the computing power of cloud connectivity, edge computing, data centers and HPC. There are any number of resources that discuss the best frameworks and…
End of content
End of content