Tensor valued initial conditions

There is also a simple functional language, which allows for the generation of tensor-valued operations. These functions take the form func $(n_1,n_2,\ldots,n_r)$ where $r$ is the desired rank, and $n_1,n_2,$ etc are the dimensions of the tensor. Available functions include:

name	description
one	the tensor is filled with `1'
zero	the tensor is filled with `0'
iota	the arithmetic sequence $(0,1,...\prod_in_i)$
eye	diagonal elements filled with `1', offdiagonal `0'
rand	tensor filled with random numbers in the range $[0,1)$

• eye is equivalent to one for vectors.
• rand generates different random numbers each time the simulation is reset, and uses the clib rand() function.