Many workloads are still CPU only. Density becomes increasingly important as cluster size grows. This scalable rack unit (SRU) focuses on density. We tested six specific applications: HPL, HPCG, STREAM, OSU, GROMACS, and WRF on the reference system in Table 5.

High Performance Linpack (HPL) is a classic HPC benchmark that has been used to measure system size and performance for many years (TOP500). The benchmark performs a LU factorization on a large matrix.

High Performance Conjugant Gradients (HPCG) is designed to complement HPL with data access and compute patterns that more closely match current workloads in HPC.

STREAM measures sustained memory bandwidth with four different kernels: Copy, Scale, Add, and Triad. Triad is the most frequently requested of these four kernels.

OSU Micro-Benchmarks are a suite of network benchmarks that can test many aspects of the network performance such as latency, bandwidth, message rate for point-to-point tests and collectives.

GROMACS is a popular molecular dynamics simulation application with a wide range of use cases. Tested with the HECBioSim 3 million atom dataset.

Weather Research and Forecasting (WRF) is a weather prediction application that is popular for both research and real-world forecasting needs. Tested with the CONUS 2.5km dataset updated to support WRFv4.