next up previous contents
Next: 8.1 An Optimization Scenario Up: Dissertation Previous: 7.6.1 The Curve Viewer


8. Technology Optimization

An existing technology can be tuned in order to achieve optimum process characteristics, and to center the process to the variation margins of the design specifications. It is obvious that such an optimum has to include multiple aspects of the process in order to obtain reasonable results. This means that as many parameters of the resulting devices as possible need to be taken into account. Moreover, multiple categories of devices must be included in the optimization procedure since otherwise a single optimized device category (and several sub-optimal ones) would be obtained instead of an optimized integrated circuit.

These requirements have two implications. For the first, this multi-aspect and multi-device simulation calls for a flexible TCAD environment giving an engineer the freedom to implement such an optimization framework. For the second, the performance of this TCAD system -- in terms of computation efficiency -- needs to be quite high, since the optimization will consume considerable amounts of resources unless special measures are taken.

By means of models SIESTA offers features which enable an efficient and maintainable implementation of the elaborate modeling network which is required for this task. Additionally, SIESTA performs the evaluation of these models in an efficient way. Parallel evaluation in concert with job-farming and load balancing on a workstation cluster reduces the simulation time to reasonable time frames. SIESTA's fault tolerant simulation-flow-models enable a smooth operation of the optimization procedure and, therefore, relieve its users from overwhelming maintenance and care.




next up previous contents
Next: 8.1 An Optimization Scenario Up: Dissertation Previous: 7.6.1 The Curve Viewer
Rudi Strasser
1999-05-27