The DOE-funded Surrogate Benchmark Initiative (SBI) is a collaborative effort involving Indiana University, UTK/ICL, and Rutgers University that aims to provide new benchmarks and tools for assessing deep neural network “surrogate” models. Trained on data produced by ensemble runs of a given HPC simulation, a surrogate model can imitate—with high fidelity—part or all of that simulation and produce the same outcomes for a given set of inputs while requiring far less time and energy.
At present, however, there are no accepted benchmarks to evaluate these surrogate models, and there is no easy way to measure progress or inform the codesign of new HPC systems to support their use. SBI aims to address this fundamental problem by creating a community repository and a Findable, Accessible, Interoperable, and Reusable (FAIR) data ecosystem for HPC application surrogate benchmarks, including data, code, and all relevant collateral artifacts that the science and engineering community needs to use and reuse these data sets and surrogates.
In Collaboration With
- Indiana University
- Rutgers University
With Support From
- The United States Department of Energy