Welcome to the April edition of our Ask an Applied Engineer series! Each month, our engineers answer your most pressing questions.
Whether you’re troubleshooting software challenges or seeking industry advice, our team is here to help you succeed.
Submit your questions here and get a response from an Applied CAx engineer.
This month, we’re answering questions about STAR-CCM+ memory and Mendix nano vs. micro flows.
Question #1
Q: Does STAR-CCM+ allow setting a limit on memory usage for meshing, and what hardware factors affect meshing performance most?
A: STAR-CCM+ does not have an option to directly limit memory usage during the meshing process. Instead, system-level controls can be activated to limit the memory STAR-CCM+ has access to. However, the limit would apply to the entire simulation, including simulating and post-processing, where local memory resources are still important.
RAM is the most critical component in the meshing process as the meshing algorithms build large in-memory data structures. If there is insufficient RAM, STAR-CCM+ must utilize hard disk memory, which significantly slows the meshing process. High-performance SSDs can mitigate this slowdown but can’t eliminate it.
Beyond memory capacity, CPUs are an important factor that can drastically affect meshing performance. Most meshing procedures have good parallel scaling, and utilizing additional CPU cores can accelerate the meshing process.
However, communication between cores can introduce additional slowdowns if too many cores are used. For meshing processes that are primarily handled serially (e.g., the Prism Layer mesher), the CPU clock speed dictates the speed at which meshing occurs.
Answered by Jaymes Dionne, PhD, Application Engineer at Applied CAx
Question #2
Q: What’s a nanoflow? When should I pick a nanoflow over a microflow?
A: Nanoflows run client-side while microflows run server-side. Use nanoflows for instant UI updates, offline functionality, and client-side validation without server roundtrips. Use microflows for database operations. Essentially, if it touches the database, use a microflow; if it’s UI level only, use a nanoflow.
Answered by Chris Ochs, Software Engineer at Applied CAx
Ask your question here to be featured in the next edition of Ask an Applied Engineer!