Simcenter STAR-CCM+ Optimizing Siemens Simcenter STAR-CCM+ Performance on AWS HPC7G EC2 Instances

When running Siemens Simcenter STAR-CCM+ on the cloud, it's essential to consider the optimal number of cells per cluster to ensure efficient performance and resource utilization. This article provides recommendations for the maximum cell count per cluster when using HPC7G EC2 instances on Amazon Web Services (AWS), which will be initially available in the US and Asia Pacific (APAC) regions.

Key Assumptions:

1. The average memory requirement is 2.4 GB per million cells. However, this value may vary depending on the physics complexity and the model splitting into smaller or larger numbers of cores.
2. The recommendations are based on the available memory of HPC7G EC2 instances.

Recommended Maximum Cell Count per Cluster Size:

The table above showcases the recommended maximum cell count for each cluster size based on the total memory available in HPC7G EC2 instances. These recommendations aim to strike a balance between computational performance and memory utilization.

When selecting the appropriate cluster size and cell count, consider the following factors:

1. Model complexity: If your model involves complex physics or requires additional memory for specific solver settings, you may need to adjust the cell count accordingly.
2. Core splitting: How you split your model across multiple cores can impact memory usage. Splitting the model into smaller or larger numbers of cores may affect the optimal cell count per cluster.
3. Scaling efficiency: While increasing the cluster size allows for a higher cell count, it's important to note that scaling efficiency may diminish as the cluster size grows. Consider the trade-off between performance gains and cost-effectiveness when choosing the cluster size.

Follow these recommendations as a starting point to optimize your Simcenter STAR-CCM+ simulations on AWS HPC7G EC2 instances. Monitor your simulations' performance and memory usage, and make adjustments based on your specific model requirements and computational objectives.

By carefully selecting the appropriate cluster size and cell count, you can maximize the performance of your Simcenter STAR-CCM+ simulations while efficiently utilizing the resources provided by AWS HPC7G EC2 instances, which will be first rolled out in the US and APAC regions.