For 15 years, ALCF has operated high performance computing resources that are used to conduct nationally and globally relevant research. In 2020, this included supporting our user community’s efforts to meet urgent challenges related to the COVID-19 pandemic. In an unprecedented pivot of its workforce and operations, the ALCF responded as needed: allocating significant computing cycles for COVID-19 research, growing its GPU capabilities to bring more computational power to addressing the pandemic, and joining a national effort to coordinate a unified computational response.
All the while, our users continued to push the boundaries of many other scientific domains. Theta delivered more than 20 million node-hours to 14 INCITE projects in 2020, advancing studies ranging from modeling supernova explosions to developing a leading-edge climate and Earth system model to exploring potential novel cancer therapies.
As ALCF prepares for the 2022 launch of its exascale system, Aurora, we continue our collaborative work through the Early Science Program and DOE’s Exascale Computing Project, along with the emerging testbeds in Argonne’s Joint Laboratory for System Evaluation (JLSE), to help science teams advance their code development efforts for research. JLSE currently operates a number of testbeds used to prepare for Aurora, alongside many others used to support more than 80 projects ranging from application portability to software development to tools and compiler development.
ALCF’s AI pathfinding efforts are allowing researchers to evaluate machine-learning-based HPC science applications running on system accelerators from a variety of vendors. This experimental hardware and software environment is available to the Argonne community now, with plans to expand to a wider audience in the near future.
Each year, ALCF continues to grow our outreach and training programs, extend our partnerships, and add new services and testbed resources. All these activities are allowing us to expand our capabilities and capacity to do science, follow the data, and employ new modes of computing.
There is still much uncertainty that lies ahead, post-pandemic, but I remain certain of the importance of our mission as a leadership computing facility and of our team’s commitment to fulfilling it.