The Advanced Photon Source (APS) and Argonne Leadership Computational Facility (ALCF) at Argonne National Laboratory offered researchers a formidable scientific and computational resource in the world. Every year, thousands of scientists employ the APS to investigate new materials, making it one of the world’s most widely used X-ray light sources.
Researchers can use the APS to observe the molecular structure of proteins down to the atomic level and to track individual ions as they flow through a battery. However, data analysis is required, and here is where the ALCF’s cutting-edge computing resources shine.
Open to researchers worldwide, the APS and ALCF are user facilities supported by the Office of Science at the Department of Energy’s (DOE) Argonne National Laboratory. About 5,500 researchers from all over the world use its ultrabright laser beams each year to examine various materials at great depth.
In many fields, from drug discovery to materials research, progress is sped up when experimental science and data analysis are combined. However, no matter how quickly all these resources can work together, faster is always better.
The APS is getting an upgrade that will make its X-ray beams up to 500 times brighter, which will hasten scientific discoveries. The ALCF is increasing the capability of its supercomputers and improving its data transport capabilities so that researchers may examine the results of their analyses in record time.
“We’ve dedicated four racks of nodes to look at the integration of experimental science and high-performance computing,” Michael Papka, Director of the ALCF and a Deputy Associate Laboratory Director at Argonne remarked.
The ALCF introduced their newest supercomputer in August 2022, which has been instrumental in reducing the time required to complete complex data analysis jobs. The result is automated, real-time data analysis that can guide scientific investigations.
The new system is being tested by Nicholas Schwarz, Principal Computer Scientist and group leader for the APS. The purpose of employing X-rays to detect the formation of microscopic fractures in a novel material is to rapidly train the X-ray instruments on the likely area where the cracks will occur next to observe the material’s behaviour.
Researchers can optimise their experiment in real-time by adjusting in response to data that arrives in seconds rather than hours. Using data from four distinct X-ray methods, Schwarz and his colleagues at the ALCF have been testing the updated APS.
However, determining when to schedule the machine’s computing time is a moving target. The ALCF group is figuring out pre-emption with Polaris, which prioritises some projects above others. The goal is near-real-time analysis for time-sensitive operations, which competes with the usual volume of massive, time-consuming tasks.
“We need to examine the available rack space to determine the best location for those jobs so that we can effectively handle the resulting conflict. Similar to a game of Tetris. The racks may be kept active while still having availability for urgent work if we plan ahead,” said Bill Allcock, Director of Operations at the ALCF.
Using information gathered from an APS experiment with no human interference, the team has recently finished their first completely automated end-to-end test of the preemptible queues on Polaris. Globus, developed by scientists at Argonne National Laboratory and the University of Chicago, automates parts of the procedure. Computational fluxes between the two institutions will be managed via Globus.
Globus coordinates the many tasks involved in an experiment, such as high-speed data transmission, ALCF computations, and data categorisation and distribution. The objective is to make this possible in all APS experiments. Researchers throughout the world will have access to ever-more-powerful computing as the APS continues its renovation and the ALCF deploys more powerful supercomputers and increased data transfer capabilities.