The batch files now output not just to one file for the product matrices Z'Y, X'Y and Y'Y but rather nvb files where nvb is the number of voxel batches to be run during the next stage. This gives a massive improvement in terms of computation speed as there was a bottleneck where each batch job was having to wait its turn to output to Z'Y for large models.
An iteration limit has been added in case voxels do not converge.
This PR includes the following updates:
nvbfiles wherenvbis the number of voxel batches to be run during the next stage. This gives a massive improvement in terms of computation speed as there was a bottleneck where each batch job was having to wait its turn to output to Z'Y for large models.