we have now a MultiVector, what is actually an array of vectors of the same format.

It allows to formulate algorithms more compact, and also allows for additional optimizations, e.g. for pair-wise inner products and simultaneous matrix vector multiplication.

A use-case for a blocked eigenvalue solver is now in the tutorial:
Model order reduction is another typical application for the MultiVector

Suggestions for additional useful features are welcome !

  • Joachim