Draft: BUG: sp model was not working properly for PFFT + resolve Issue #228

The stochastic plasticity model was not working correct when using PFFT, because it made some assumptions that were hard coded for a FFTW decomposition of the domain. It works now properly for FFTW and PFFT. The PFFT bug fixes are tested vs the FFTW implementation. I introduced therefore a FFTEngineType flag in the FFTEngineBase the flag can tell i.e. the cell which FFT is used. By this it is possible to tell the stochastic_plasticity_material which decomposition is applied. I also fixed the BUG of Issue #228.