by Karl Freund | Jun 30, 2025
While GPU performance has been the focus in data centers over the last few years, the performance of fabrics has become a key enabler or bottleneck in achieving the throughput and latency required to create and deliver artificial intelligence at scale. Cambrian-AI...
by Karl Freund | Mar 4, 2025
Training vs. Inference: Diverging Compute Demands in Conversational AI This article was written by ChatGPT’s new Research offering. It took 6 minutes of compute and researched 42 articles. My intent was to both shed some light on the topic, and on ChatGPT’s new...
by Karl Freund | Oct 24, 2024
Many companies, including Nvidia, use Digital Twins to model data centers, helping to avoid waste and hotspots as data centers move to an AI-capable data center. Karl Freund and Dr. Jonathan Koomey collaborated to produce this paper that researches the practice and...
by Karl Freund | Oct 15, 2024
During the last few years, we have witnessed explosive growth in the data center,driven by the pervasive use of public cloud services, big data analytics, and most recently by the development of Artificial Intelligence (AI) enabled by power and data-hungry GPUs. This...
by Karl Freund | Sep 26, 2023
Last year, IBM launched the z16 with an integrated AI accelerator on each CPU chip. Now, with the infusion of AI into IBM z/OS and a robust AI open-source toolkit, IBM Z customers can realize low-latency AI on a highly trustworthy and secure enterprise system: the...