Products: Abaqus/Standard Abaqus/AMS
Benefits: The performance of frequency extraction analysis using the AMS eigensolver has been improved for large models.
Description: For large models frequency extraction analyses using the AMS eigensolver are enhanced as follows:
Parallel performance enhancement of the reduction phase in the AMS eigensolver
Performance improvement of multipoint constraint recovery for selective recovery of eigenmodes using the AMS eigensolver
Performance improvement for transforming the acoustic-structural coupling operator for acoustic-structural coupling analysis using the AMS eigensolver
Table 4–1 illustrates the performance improvements for frequency extraction analyses using the AMS eigensolver for two industrial models: Model 1 is a 22 million degree-of-freedom vehicle body model, and Model 2 is a 20 million degree-of-freedom vehicle body model. All analyses were run on a Linux machine with dual 2.8 GHz Intel Ivy Bridge processors (2 × 10 cores), 192 GB memory. In all cases, all 20 cores are used for the computations.
Table 4–1 Performance improvements for the frequency extraction analysis using the AMS eigensolver.
Model | Degrees of Freedom | Number of Modes | Abaqus 6.14 | Abaqus 2016 | Speedup Factor (Abaqus 6.14/2016) | ||
---|---|---|---|---|---|---|---|
Run Time (sec.) | Memory Usage (GB) | Run Time (sec.) | Memory Usage (GB) | ||||
Model 1 | 22 million | 8,615 | 4778 | 169 | 3107 | 167 | 1.54 |
Model 2 | 20 million | 12,259 | 2819 | 143 | 2290 | 116 | 1.23 |