CEETRON Solve is a multitask SDK that accelerates the implementation of CAE simulation applications by providing a developer-friendly framework to both define and solve FEM models. It is a FEM toolkit designed for both new and existing solvers. When building a FEM application from scratch or seeking to speed up an existing one, the framework and linear algebra algorithms in CEETRON Solve provide a toolbox of technologies to fit the CAE developer’s needs. Whether or not you rely on Ceetron Solve to build your local matrices, you can trust its SMP direct and iterative linear solvers as well as its eigensolvers to deliver the CPU performance you need. CEETRON Solve works behind the scenes of major industry-standard CAE packages, providing established performance on shared-memory parallel architectures. The solvers are built on a high level of coarse-grained data distribution and rely on an extensive use of optimized BLAS for low-level operations. However, the true difference is in data management choices, multithreading and memory optimization, and a continuous quest to trim off CPU.

Features:

• Element-Level Contributions: Off-the-shelf calculation of local matrices and vectors for standard structural and thermal analysis, connected to a broad and user-expandable material model library.
• Parallel Direct Linear Solver: Proprietary high-performance SMP multifrontal sparse linear solving, including optimized functions for multiple right-hand side resolution on the same matrix and threaded I/O for out-of-core operation.
• Parallel Iterative Solvers: Proprietary iterative preconditioned conjugate gradient and GMRES solvers to tackle both symmetric and non-symmetric systems.
• Parallel Eigensolvers: Proprietary eigensolvers to address specific eigenvalue problems: Lanczos or Automated Multilevel Substructuring to extract normal modes of vibration, and a subspace algorithm generally used to determine buckling modes.