Durham, NC, May 28, 2024 — As the adoption of artificial intelligence (AI) has grown, so has the popularity of graphical processing units (GPUs) and cloud-based platforms for supporting AI workloads. While these platforms may be necessary for efficiently running many AI model training and inferencing functions, other parts of the AI development workflow can run well on CPU-only workstations. In fact, Intel has designed hardware and software specifically for these kinds of AI tasks.
To explore the potential of these alternative platforms for parts of the AI development workflow, Principled Technologies performed proof-of-concept testing. They carried out three AI development workflows using only the CPU cores in tower and mobile workstations featuring Intel processors. The workflows involved (1) characterizing documents then adding them to a database and indexing them, (2) analyzing a portrait by asking a local large language model to describe it, and (3) standardizing images and creating new images suitable for testing or training.
As the report concludes, “For all three AI workflows we executed, we consider the time the workstations needed to complete the tasks to be acceptable, and believe that these workstations can be appropriate, cost-effective choices for these kinds of activities.”