A couple of changes to the matrix.py module. The ODE solver gave me the idea of how important it was to cut down on copy constructors and destructors. So there were two functions in the Matrix class that would come under that category and will be used almost everywhere.
The function added has been the "resize" function. So if a matrix exists but is now equated to another matrix of another dimension, the object will be modified rather than deleted and built again.
And the zeros function which would initialize self.data in the beginning and then rebuild it now will check whether the object exists. If it doesn't exist, it will initialize and fill it. If the new dimension is different from the existing dimensions, it will resize it.
The __init__ function calls the function zeros with the express arguments of the actual rows and columns of the matrix different from default rows and columns so that a resize takes place.
Finally, the __call__ function has been changed to check if the dimensions are different and resize if so otherwise just equate them.
The addition, subtraction and multiplication operations still need a constructor for the result and the functions will then return that result. But I will try to reduce the need for those in simulations but keep them for error checking as it makes it much easier to type out these operators as regular commands.
The function added has been the "resize" function. So if a matrix exists but is now equated to another matrix of another dimension, the object will be modified rather than deleted and built again.
And the zeros function which would initialize self.data in the beginning and then rebuild it now will check whether the object exists. If it doesn't exist, it will initialize and fill it. If the new dimension is different from the existing dimensions, it will resize it.
The __init__ function calls the function zeros with the express arguments of the actual rows and columns of the matrix different from default rows and columns so that a resize takes place.
Finally, the __call__ function has been changed to check if the dimensions are different and resize if so otherwise just equate them.
The addition, subtraction and multiplication operations still need a constructor for the result and the functions will then return that result. But I will try to reduce the need for those in simulations but keep them for error checking as it makes it much easier to type out these operators as regular commands.
No comments:
Post a Comment