September 15, 2021 – A project involving researchers from Lawrence Livermore National Laboratory (LLNL) and National Laboratories and academic collaborators has received funding from the US Department of Energy (DOE) as part of an effort to adapt the next-generation high-performance scientific software. computer systems (HPC).
The project, “ComPort: Rigorous Testing Methods to Safeguard Software Porting”, will address one of the main challenges of scientific computing: the digital aspects of porting scientific applications to different HPC platforms. The need for solutions is becoming more urgent as supercomputers increasingly integrate various combinations of central processing units and graphics processing (CPU and GPU), accelerators and software, according to the researchers.
The increasing heterogeneity of HPC hardware can affect the digital integrity of codes, which means that software running on one system may not produce the same results on another, explained Ignacio Laguna, co-principal investigator of the ComPort project and LLNL IT specialist. The challenge will become even more complex as the DOE enters the exascale era, where computers will exceed a quintillion calculations per second, Laguna added.
“The ComPort software tools will enable LLNL’s large science simulations to produce more robust and reproducible numerical calculations in next-generation supercomputers,” Laguna said.
The project will compare validated results from previous versions and rigorously test the numerical behavior of hardware and software to verify whether the calculation results match the expected responses, according to Laguna. Additionally, ComPort automated software tools will provide users with high level feedback to diagnose and repair software applications to maintain accuracy despite hardware and compiler changes.
A number of tools developed under the Lab’s Advanced Technology Development and Mitigation Next Generation Computing Enablement project, such as FLiT and FPChecker, were critical in selecting the award, Laguna said.
The project is a collaboration between LLNL, University of Utah (Principal Investigator Ganesh Gopalakrishnan and Pavel Panchekha), Pacific Northwest National Laboratory (Ang Li), University of California, Davis (Cindy Rubio-Gonzalez) and the University of Washington (Zachary Tatlock).
Recent DOE awards, totaling over $ 13 million, are administered by the DOE Office of Science, Advanced Scientific Computing Research (ASCR) program.
For a full list of funded projects, click here.
Source: Lawrence Livermore National Laboratory