Parallel RCM Numberer

The Parallel Reverse Cuthill-McKee (Parallel RCM) numberer is a parallelized version of the RCM algorithm that provides improved performance for large-scale problems while still reducing matrix bandwidth.

Description

The Parallel RCM numberer: - Uses a parallel implementation of the RCM algorithm - Distributes the reordering computation across multiple processors - Reduces setup time for large problems - Maintains the bandwidth reduction benefits of RCM - Implemented as a singleton to ensure consistent numbering across the model

Parameters

Parameter

Type

Description

None

This numberer doesn’t take any parameters

Usage Example

# Create a Femora instance


# Get a Parallel RCM numberer
fm.analysis.numberer.get_numberer("ParallelRCM")

Notes

  • Combines the benefits of RCM with parallel processing

  • Particularly effective for very large problems

  • Can significantly reduce setup time compared to serial RCM

  • Requires parallel processing capabilities

  • Recommended for large-scale problems on multi-core systems

  • Maintains the same quality of bandwidth reduction as serial RCM