Supercomputers or high performance computers (HPC) are indispensable in various industrial sectors, from computing weather forecasts to designing safer cars. Given their importance, there is a need to benchmark or evaluate HPC platforms to determine how they react to stressors such as high workloads, large numerical calculations or intense data transfer.
About 20% of supercomputing systems are used to derive properties of elementary particles using lattice gauge theory (LGT), which therefore provides one of the most direct ways to evaluate the performance of supercomputers. Prof Biagio Lucini’s group at Swansea applied their expertise in Beyond the Standard Model physics to develop BSMBench, a code based on LGT methods. The unique property of the code is the capacity to vary computational and communication aspects simply by adjusting basic parameters of the theories it simulates. This makes it a powerful benchmarking tool for supercomputer systems.
BSMBench has been used by a number of supercomputing companies (NVIDIA, Intel) to benchmark their new hardware products. It has also been used to demonstrate the effectiveness of supercomputing system after securing them against malware, which is important for customer confidence. Developed by Ed Bennett, BSMBench was released as an open source project in June 2012 and can be downloaded free of charge from https://gitlab.com/edbennett/BSMBench.