Pioneering supercomputer will enable scientists and engineers to optimize hardware to support groundbreaking research
Backed by a multi-million dollar federal grant, a research team from three major universities will soon begin work on a pioneering supercomputing system that allows scientists and engineers to align its processors, accelerators, memory and other hardware components to better meet their needs.
This innovative system will leverage increasingly complex levels of software while avoiding the hardware bottlenecks that often hamper high-level calculations. This system will allow researchers to perform calculations and solve problems that today’s supercomputers cannot handle.
On October 1, 2021, researchers from Texas A&M University, University of Illinois Urbana-Champaign (UIUC) and Texas Advanced Computing Center (TACC) at the University of Texas at Austin (UT Austin) will begin to collaborate on a prototype of what they call the Accelerating Computing for Emerging Sciences (ACES) system.
The National Science Foundation (NSF) will provide $ 5 million for the development of ACES and an additional $ 1 million per year over five years to pay for the operation and support of the system.
Texas A&M Acting Vice President for Research Jack Baldaulf expressed gratitude to the NSF for its substantial investment in the ACES project. “We are grateful to the NSF for the opportunity to lead such an important initiative and to our staff and collaborators from Texas A&M HPRC at UT Austin and UIUC for making this effort a success,” said Baldauf. . “Computer science is essential to our national needs and the ACES platform will not only advance research, but will also help train the future workforce in this field. “
The team’s goal is to develop a comprehensive system that will serve researchers in a wide range of academic disciplines and computing skills, according to Honggao Liu, executive director of High Performance Research Computing (HPRC) at Texas A&M and principal investigator of Texas A&M. project.
These disciplines include artificial intelligence and machine learning, cybersecurity, health population informatics, genomics and bioinformatics, human and agricultural life sciences, oil and gas simulations, design of new materials. , climate modeling, molecular dynamics, quantum computing architectures, imagery, smart and connected societies, geosciences and quantum chemistry.
“The ACES system will support the national research community through coordination systems supported by the NSF,” Liu said. “In this way, the ACES system will provide invaluable support to cutting-edge projects in a wide range of research disciplines in the country. ACES will also leverage HPRC’s efforts that promote science and expand participation in computing at the K-12, college and professional levels to have a transformative impact nationwide by focusing on training, education and awareness.
Researchers should think of ACES as a cyber buffet, said Timothy M. Cockerill, director of user services, TACC at UT Austin, and co-principal investigator of the ACES project. “They will essentially be able to create the custom environment they need for each task and will not be limited to the content of a physical server node,” said Cockerill.
ACES will open new avenues for scientific progress, said Shaowen Wang, head of the Department of Geography and Geographic Information Sciences, professor at UIUC and co-principal investigator of the ACES project. “Exciting advancements across many scientific frontiers will become possible by harnessing the hybrid computing resources and highly adaptable framework offered by ACES to enable increasingly complex science workflows driven by geospatial big data and artificial intelligence,” said Wang said.
Lisa M. Perez, associate director of advanced IT activation, and Dhruva Chakravorty, associate director of user services and research, both of the HPRC at Texas A&M, are also co-principal investigators.
Research that generates groundbreaking discoveries will require highly advanced computer designs that can meet the challenge, said Costas N. Georghiades, senior associate vice president for research at Texas A&M. “With the increasing complexity of computational issues in the big data age we live in, it’s no longer enough to use traditional supercomputers that rely solely on software optimization,” Georghiades said. “The ACES system will also be able to scale hardware resources on the fly to tackle complex computational tasks more effectively. Texas A&M is proud to lead this effort in conjunction with our university partners at UT Austin and Illinois.
ACES leverages an innovative composable framework via PCIe (Peripheral Component Interconnect Express) Gen5 on Intel’s upcoming Sapphire Rapid (SPR) processors to deliver a rich accelerator test bed comprised of Intel Ponte Vecchio (PVC) GPUs (graphics processing units), Intel FPGA (Field Programmable Gate Arrays), NEC vector engines, NextSilicon coprocessors and Graphcore Intelligence Processing Units (IPU).
The accelerators pair with Intel Optane memory and DDN Luster storage interconnected with Mellanox NDR 400 Gbps InfiniBand to support workflows that benefit from optimized devices. ACES will enable applications and workflows to dynamically integrate different accelerators, memories and networked computing protocols to glean new information by rapidly processing large volumes of data and will provide researchers with a single platform to produce models of complex hybrid programming to effectively support calculations that were not feasible before.
About the research at Texas A&M University: As one of the world’s leading research institutes, Texas A&M is at the forefront of making significant contributions to scholarship and discovery, including science and technology. Research conducted at Texas A&M generated more than $ 1.131 billion in annual expenses in fiscal 2020. Texas A&M ranked in the top 25 in the most recent research and development survey. National Science Foundation higher education based on spending of over $ 952 million in fiscal 2019 Texas A&M research creates new knowledge that provides foundational, foundational and applied contributions resulting in economic benefits for the state, the nation and the world. research.tamu.edu