Intel Processor Selection for Ansys Mechanical and Fluent Workloads

To assist customers, Intel collaborated with Ansys and with MVConcept to measure the performance of select Intel Xeon Scalable processor families using both single and multi-node Ansys Fluent and Ansys Mechanical official benchmarks spanning a range of size regimes and solvers. The performance results reflect the complex relationship between computational efficiency, processor core count, core frequency and memory bandwidth.

Ansys workloads have both memory bandwidth and compute intensive requirements which can vary for many reasons, including dataset size and the solver utilized. High performance requires balancing the efficiency of the processor core against the bandwidth of the memory subsystem. Given the one process per core nature of Ansys workloads and core-utilization basis for Ansys licenses, performance per core is the key metric in benchmark analysis as it reflects the expected performance in the field, but only in situations where there is sufficient memory bandwidth to keep the computational units on the processor cores supplied with data and running efficiently.

This white paper presents and evaluates the performance behavior of 2nd Generation Intel Xeon Scalable processors from the Intel Xeon Gold 62xx and Intel Xeon Platinum 92xx families of processors.



Request Free!