Center for High Throughput Computing (CHTC)
About this resource
Established in 2006, the Center for High Throughput Computing (CHTC) is committed to democratizing access to powerful computing resources across all research domains. High Throughput Computing (HTC) encompasses a set of principles and techniques designed to optimize computing resource utilization towards solving complex problems. When applied to scientific computing, HTC enhances resource efficiency, automation, and accelerates scientific breakthroughs, including those in machine learning.
Are you a researcher at UW-Madison seeking to extend your computing capabilities beyond local resources, particularly for machine learning tasks? Request an account now to take advantage of the open computing services offered by the CHTC!
CHTC “Recipes”
Visit the CHTC Recipes Repository to discover a collection of common CHTC workflows or “recipes”, including those specifically geared towards machine learning tasks.
GPU-based
Explore our collection of templates tailored for high throughput compute (HTC) systems utilizing GPUs, ideal for accelerating machine learning workflows. These templates streamline the process of job submission, maximizing the utilization of GPU resources for your computational tasks in machine learning. Dive into efficient computing with our GPU-based templates available on GitHub: CHTC GPU Templates
Container Guides
Empower your machine learning research endeavors with containerization! CHTC’s guides on Docker and Apptainer for HTC empower researchers to encapsulate their machine learning workflows, dependencies, and environments efficiently. Seamlessly integrate containers into your machine learning computing workflow for enhanced reproducibility and scalability.