PREA
PREA
Activates resource prediction for supported analysis and solver types.
Notes
Note: The PREA command requires both an Ansys Mechanical Enterprise license and a Structures AI+ license.
Use the PREA command to begin resource prediction for the current load step, analysis type, solver type, and core count. This command estimates the following:
Solution elapsed time
Total memory usage
Total disk space usage
GPU recommendation
Usage
Issue PREA after you set the analysis type and solver type to ensure accurate prediction.
A successful prediction does not guarantee a successful solution. For example, resource prediction cannot determine convergence for nonlinear static analyses or PCG iterative solvers.
Predictions apply only to solutions using distributed-memory parallel (DMP) processing. Results do not reflect hybrid parallel usage.
The resource prediction feature generates two charts:
First chart: Shows predictions for the core count specified with the
-npoption. This chart can include any number of cores. For core counts above 128, PREA still generates predictions, but accuracy diminishes.Second chart: Shows predictions for a range of core counts from 4 to 128.
Supported Analysis and Solver Types
Predictions for solution time, memory, and disk usage:
Static analysis with sparse direct equation solver
Static analysis with PCG iterative solver
Modal analysis with Block Lanczos solver using sparse direct equation solver
Modal analysis with PCG Lanczos solver
Harmonic analysis using the Full method with frequency or mesh decomposition
Harmonic analysis using the VT method with frequency or mesh decomposition
Predictions for GPU recommendation:
Static linear and nonlinear analysis with direct or iterative solver
Modal analysis with LANB or LANPCG solver
Harmonic analysis using the full method (FULL) with mesh decomposition
Predicted Metrics and Output Format
| Metric | Output Format |
|---|---|
| Solution Elapsed Time | Displayed as a table in the output |
| Total Memory Used | Displayed as a table in the output |
| Total Disk Space Used | Displayed as a table in the output |
| GPU Recommendation | Displayed as a table in the output |
GPU Accelerator Requirements
To maximize GPU accelerator performance, see Requirements for the GPU Accelerator in the Mechanical APDL Application (Linux) and Requirements for the GPU Accelerator in the Mechanical APDL Application in the Ansys, Inc. Installation Guides (Windows).
Resource Prediction in the Mechanical Application
For information on using resource prediction in the Mechanical application, see Understanding Resource Prediction for the Solution in the Mechanical User's Guide.