AI is accelerating at a pace few industries have ever witnessed. Every six months, model sizes are doubling, workloads are scaling across thousands of GPUs, and the demands on compute infrastructure are surging beyond anything imagined even a few years ago.

Servers are challenged to process more intricate and diverse types of workloads in cloud, hybrid-cloud and enterprise data centers.

Advanced IP Design for PCI Express^® Technology

By Scott Knowlton, PCI-SIG MWG Co-Chair

With the widespread adoption of compute-intensive workloads – such as artificial intelligence and machine learning – in enterprise and cloud data centers, high-speed, low-latency interconnects like PCI Express^® architecture are required to connect high-performance nodes.

PCI-SIG has built its reputation on delivering high quality PCI Express^® (PCIe) specifications that have doubled bandwidth on average every three years, while maintaining full backwards compatibility with prior generations.