6. Code generation for computations

Warning

This section of the documentation is still incomplete.

Computation trees

Sets of computations organized as dependency graphs, with multiple roots (input variables) and leaves (output expressions).

6.1. Defining a computation

class csparse
cs=csparse()

Creates an empty computation tree

Returns

empty csparse object

declareSet(cs, destination, functionName)

Adds an input variable to the tree

Parameters

  • cs – csparse object

  • destination – Tcalculus tensor variable created using Tvariable()

  • functionname (string) – name of the function to be created

declareGet(cs, sources, functionName)

Adds output expressions to the tree

Parameters

  • cs – csparse object

  • sources – cell array of Tcalculus tensor-valued expressions

  • functionname (string) – name of the function to be created

declareCopy(cs, destinations, sources, functionName)

Sets the value of input variables, based on the value of output expressions

Parameters

  • cs – csparse object

  • sources – cell array of Tcalculus tensor-valued expressions

  • destinations – cell array of Tcalculus tensor variables created using Tvariable()

  • functionname (string) – name of the function to be created

6.2. Code generation

cmex2compute(parameter1, value1, parameter2, values2, ...)
class2compute(parameter1, value1, parameter2, values2, ...)

Generate code to set values of input variables, get values of output expressions, and copy output expressions to input variables.

Parameters

  • parameter1 (string) – parameter to set

  • parameter2 (string) – parameter to set, …

Returns

name of the MATLAB© class created

Return type

string