It’s representative of realworld simulation workloads rather than something contrived or less than useful like some of the OpenFOAM benchmarks Phoronix Test Suite uses.
I am primairly throwing shade at the mesh time benchmarks and the impossibly low DoF benchmarks they include.