Data visualization for CSE-based simulations is a fundamental capability as it enables to analyze the simulation results and thus allows to ultimately derive conclusions, further driving the research. Therefore, data visualization capabilities are a central aspect of an interactive simulation framework. The diversity of simulations requires flexible visualization tools, such as three-dimensional rendering methods supporting, for instance, scalar- or vector-field visualization on top of a mesh. Especially important is to give the individual framework modules full access to the rendering backend, as the visualization demands for all future modules are not foreseeable. For instance, simulation tools might require a simulation-specific visualization especially for three-dimensional data, such as a combination of rendering algorithms to reveal the behavior inside of the simulation object.
Also, convenient selection of a specific quantity to be rendered is required as well as different mesh representations, e.g., wireframe3 . The latter can be coupled with automatic mesh evaluation algorithms aiding an advanced user, referring to a user with an additional background in mesh generation, in the task of evaluating the mesh quality. These additional investigation methods further underline the importance of a modular simulation environment, capable of attaching non-simulation modules, such as mesh generation and evaluation modules, to the simulation platform.
Aside from supporting three-dimensional rendering, support for chart visualization is essential. Chart visualizations enable to investigate a set of quantities relative to a reference quantity. For example, a current-voltage characteristics enables to judge a device’s performance. Vital to such a mechanism is the ability to access data generated from several simulation runs to, for instance, compare a set of current-voltage characteristics for the same device but for different doping profiles. The data from the individual simulation runs should be visualized in the same chart, enabling the end user to relate the quantities to each other.
In general, essential to visualization is the ability for comparing results. For instance, two different charts are compared or two different render views of the same mesh are displayed simultaneously. This requires the visualization backend to support an arbitrary number of visualization windows, in an arbitrary combination. Exemplary combinations would be three rendering windows and two chart windows as well as two rendering windows and four chart windows. Related to this mechanism is the ability to analyze data in a multi-monitor environment. Today’s typical workstation setups are increasingly equipped with more than one monitor, introducing the ability to utilize this setup for visualization needs. For instance, where two chart views are displayed on one monitor, one rendering view is displayed on the second. While this requirement seems trivial, it introduces substantial design considerations concerning the general GUI platform of an interactive simulation framework and is thus vital to be explicitly stated as a requirement.
Simulation tools may generate a series of simulation results which has to be supported by the visualization backend. For instance, a device simulator performs a range of simulations according to a range of contact values, allowing to determine the current-voltage characteristics of a device. In such a case, the visualization backend must support manual or automatic stepping through the simulation results. Where the first allows to specifically analyze results, the latter enables a visualization in a movie-like manner, further improving the perception of the simulated physical processes.