Plasticity is a phenomenon, which occurs in many materials, for instance wood, concrete, metals, asphalt and snow. Many important properties of the material model are encoded in the yield surface, which is the sub-manifold of the principal stress space where the yield function equals zero. In the contribution at hand, an algorithm is detailed, which can identify arbitrary yield surfaces. No assumptions regarding linearity of the material, kinematics or isotropy are invoked [1]. Furthermore, the applicability on high performance computers is discussed. The continuum mechanical fundamentals of the algorithm are briefly summarized. Furthermore, it is outlined how the proposed algorithm can be applied to crystal plasticity as well. Afterwards, numerical examples for verification purposes are demonstrated. Furthermore, the applicability to multiscale analyses and anisotropy are demonstrated by numerical examples. The numerical examples include FE 2 and analytically solvable representative volume elements. Based upon these multiscale frameworks, the influence of material parameters on the microscale onto the macroscopic yield surface are studied. The proposed algorithm is suited to generate training data for the application of neuronal networks [1, 2].