Changes between Initial Version and Version 1 of Ticket #6236, comment 8


Ignore:
Timestamp:
2020-11-23T10:14:03Z (4 years ago)
Author:
Philip Hannebohm

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #6236, comment 8

    initial v1  
    1010> I'm not sure if we really want to go above 2x2. Maybe it would make sense for sparse equations, e.g. if there are no more than 2N nonzero elements in a NxN system, otherwise the complexity will just blow up very quickly and it's very unlikely you'll get anything simple out of that.
    1111
    12 This is a good question. For sparse systems there aren't as many determinants to be built. There may be an optimal sparsity and size NxN, for which this is still faster than a numerical solver.
     12This is a good question. For sparse systems there aren't as many determinants to be built. There may be an optimal sparsity and size NxN, for which this is still faster than a numerical solver. Also for the `Di`s there are a lot of sub-determinants that can be computed only once.