Add a keyword in run.in to compute the elastic constant tensor about gpumd

[brucefan1983]

Now cubic crystal is done. I may do hexagonal crystal next.

[brucefan1983]

Perhaps a more elegant method of computing the elastic constants (at any temperature) is to use strain-fluctuations. Penhua Ying has written a nice tutorial here:

https://github.com/hityingph/Tutorial-on-atomic-simulations/tree/main/2.%20Elastic%20Constants%20Based%20on%20Strain%20Fluctuation%20Method

[brucefan1983]

So I suggest using the method as in Penghua Ying's tutorial. I don't want to implement the zero-temperature approach for all the crystals any more. Too tedious.