Numerical Mathematics and Scientific Computing

Research Group

Software

BOP - A Simulator for Large-Scale Process Engineering Problems
The simulator BOP (Block Orientend Process Simulator) is a software package for steady-state, transient and Monte Carlo simulation of large-scale problems from process engineering. The simulation concept is based on a divide-and-conquer strategy which is efficiently applicable on parallel computers with shared memory.
 
gltools for interactive or non-interactive graphical output
Enhancing the OpenGL API with additional functionalities, gltools has been designed with the needs of numerical analysts in mind. Thus, unlike many other packages available, it can be used to enhance existing codes with interactive or non-interactive graphical output.
 
pdelib - a FVM and FEM toolbox for PDEs
The toolbox pdelib has been developed for solving partial differential equations in one, two and three dimensions. It includes a grid generator, iterative and direct solvers for systems of equations, an online graphics package and interfaces to packages of other institutes and companies.
 
WIAS-TeSCA for two and three dimensional semi-conductor analysis
Augmented by a vast variety of additional physical phenomena playing a role in the operation of specialized semiconductor devices, WIAS-TeSCA serves to simulate numerically the charge carrier transport in semiconductor devices based upon the drift-diffusion model.
 
TetGen - a Delaunay Tetrahedral Mesh Generator
TetGen is a program for generating tetrahedral meshes from three-dimensional domains. One of its main applications is to generate suitable meshes for numerically solving PDEs by finite element and finite volume methods. TetGen uses Delaunay-based algorithms which have theoretical guarantees on mesh quality and mesh size.
 

Supported Software

MooNMD Mathematics and object-oriented Numerics
MooNMD solves equations from Computational Fluid Mechanics, as convection-diffusion equations, incompressible Navier-Stokes equations and systems consisting of these types of equations, on the basis of finite element discretizations and implicit time stepping schemes.
 
WiasPeb - numerical simulation of the Post-Exposure Bake in photolithography
WiasPeb is a python module for Post-Exposure Bake simulations. It can be used as an plugin of the Fraunhofer IISB lithographie simulator Dr. LiTHO or as an standalone package. The advandage of our module is the simulation of effects between resist region e.g. interactions between litho 2 and litho 1.
 
PARDISO PArallel spaRse DIrect SOlver
PARDISO is an OpenMP parallel direct solver for sparse matrices reaching BLAS3 performance. It includes algorithms for solving linear systems with symmetric positive definite, indefinite, Hermitian definite and indefinite, complex symmetric, structural symmetric and unsymmetric matrices. The project started at ETH Zurich, resulting in the PH.D. thesis of O. Schenk (2000). It continued at University Basel in cooperation with K. Gärtner (ETHZ, WIAS since 1998). As a new feature, version 4 incorporates incomplete factorization, thus allowing PARDISO to be used as a preconditioner in iterative methods. Partial problems are incompletely factorized recursively taking approximate estimates of norms of inverse matrices into account. The direct factorization of a much smaller remainder is the last step in that process.