LLNL gets supercomputer from AMD and Penguin Computing to aid COVID-19 research – The new agreement provides AMD for the Corona National Laboratory’s supercomputing cluster, which is expected to almost twice the peak computing power of the system, with upgraded graphics accelerators. Scientists will use the system via the Public / Private COVID-19 HPC Consortium and researchers from LLNL who are detecting potential SARS-CoV-2 virus antibodies and anti-virus compounds, which is a COVID-19 virus.
As the COVID-19 pandemic continuously affects millions of people around the world, LLNL is committed to implementing the nation’s most powerful superstructures and expertise in computer modelling and data science to fight the fatal disease.
In order to support this endeavour, Laboratory, AMD and Penguin Computing have agreed to upgrade the high-performance corona cluster (HPC) with the in-kind contribution of advanced AMD InstinctTM accelerators, which is projected to almost double the peak machine performance.
Under this agreement, AMD is delivering a Corona system’s Radeon Instinct MI50 accelerator to a speed point operating at a speed of over 4.5 petaFLOPS. It will be used by a national public / private partnership of COVID-19 HPC Consortium that offers free computing time and resources to countrywide scientific researchers engaged in the fight against COVID-19, and LLNL researchers, who are working on the detection of potential antibodies to viruses causing COVID-19 for SARS-CoV-2 virus.
The Corona system uses the Tundra Extreme Tuner platform from Penguin Computing. It supports the AMD EPYC CPUs that work side-by-side with the AMD Radeon Instinct Accelerators.
Corona’s system — named for the total 2017 solar eclipse — was delivered to LLNL in 2018 on a contract with Penguin Computing for open science applications that were not classified. AMD intends to support research into COVID-19 and to promote its partnership and cooperation with LLNL in software and tools development, but the upgrade does not cost the National Nuclear Security Administration (NNSA) at no charge.
In return for upgraded GPU, AMD ensures computing cycles that are being used for a range of purposes, such as time for LLNL COVID-19 research and proposals approved by COVID-19 HPC Consortium, and to help AMD software engineers and application specialists in the development of their software projects.
Michel McCoy, LLNL’s Advanced Simulation & Computer Director said “AMD is well known to be the key partner of the upcoming NNSA exascale-class system, the Hewlett Packard Enterprise El Capitan supercomputer.
“But a lasting partnership involves multiple partnerships with common objectives. We are pleased that AMD has provided this generous range of products, especially in view of the need for a determined pace to mitigate and, in the end, to break this pathogen. With a high bandwidth memory of up to 26.5 teraFLOPS or 13.3 teraFLOPS, the AMD accelerators delivers a floating-point performance of single-precision peaks.
“The best and brightest minds are required to work together to achieve an effective COVID-19 response. We can accelerate critical modelling and research to help combat the virus, through the massive data capabilities of the world’s most powerful supercomputers, “said Forrest Norrod, Senior Vice President and General Manager, AMD Datacenter and Embedded Systems Group. AMD is proud to support this effort, through the COVID-19 HPC Consortium and the Lawrence Livermore National Laboratory, with the contributions of processors that are now adapted to the science.
“Penguin Computing is committed to supporting international research to combat the COVID-19 virus,” said Penguin Computing Chairman Sid Mair. “Increasing Corona capabilities in HPC and AI computing could significantly improve COVID-19 research, possible vaccines, treatment and pathways of contagion.”
Corona is one of the seven LLNL supercomputers most able to access researchers through the COVID-19 HPC Consortium, which includes over one dozen government, industry and academic institutions, led by the White House Science and Technology Policy Bureau, USA. Energy and IBM Department. The consortium aims at speeding up the development of COVID-19 detection methods and therapies. On April 6, AMD formally entered the consortium.
AMD software engineers support the porting and optimization of GPU performance on relevant applications for certain applications that are essential to Corona’s COVID-19 endeavour.
The Corona system will also help LLNL scientists to pursue antibodies and antiviral medicines that could fight against the virus. COVID-19 is becoming a top priority for the Corona system, where antibody candidates for SARS-CoV-2 are virtually monitored, designed and validated and small molecules interacted with the proteins of the virus to detect antiviral compounds.
The upgrade will allow LLNL researchers to speed up the modeling of molecular interactions vital to the effort and run a wider and more diverse set of applications on the system.
“The addition of these new state-of-the-art GPUs on Corona will boost the capability of the teams working on COVID-19,” said Jim Brase, LLNL’s deputy associate director for Programs. “It’s going to allow us to go faster, with more throughput.
We’ll have more resources, so we can run more cases and potentially get to new designs for both antibodies and small molecules faster, that may lead to better treatments. They’ll also enable some of our new software, both for simulation and machine learning applications, to run more efficiently and better.”
A team of LLNL researchers has used the Corona system to evaluate therapeutic antibody designs with the use of a first-of-kind virtual screening platform, combining experimental data with machine learning, structural biology, bio-informatic and high-fidelity molecular simulations. The team has reduced the list of antibody candidates to approximately 20 different options and has started exploring other antimicrobial designs.
The researchers are convinced that this upgrade doubles the number of computer-priced simulations they perform so they will find an effective antibody design more likely.
LLNL computer scientists and computer biologists also use the Corona system to investigate thousands of small molecules with anti-virus SARS-CoV-2 properties. Increasing Corona’s speed and performance will allow researchers to carry out further, highly detailed molecular dynamics calculations in order to assess possible objectives of SARS-CoV-2 for small molecule inhibitors that could prevent or treat COVID-19 infections.
For more information on LLNL’s COVID-19 research, visit the web.
LLNL gets supercomputer from AMD and Penguin Computing to aid COVID-19 research – Stay with us for the latest update.