Source code for fracture simulaiton, based on cohesive zone model and implicit finite element method. The formulation and results are described in the following papers (under review):

• Parallel Implementation of Implicit Finite Element Model with Cohesive Zones and Collision Response using CUDA
• Cohesive Zone Micromechanical Model for Compressive and Tensile Failure of Polycrystalline Ice

This code requires Intel Math Kernel Library

Required libraries: Intel MKL, OpenTK, OpenTK.GLControl, ManagedCUDA, CUDA Toolkit 9.1. It is possible to build the library with other versions of CUDA Toolkit, but the supplied project is for CUDA 9.1. Use 64 bit architecture consistently for all libraries and PTX.

Copy MKL libraries (64 bit) to the project folder or make them accessible via PATH variable: mkl_avx.dll mkl_avx2.dll mkl_avx512.dll mkl_avx512_mic.dll mkl_core.dll mkl_def.dll mkl_intel_thread.dll mkl_mc.dll mkl_mc3.dll mkl_rt.dll mkl_sequential.dll mkl_tbb_thread.dll mkl_vml_avx.dll mkl_vml_avx2.dll mkl_vml_avx512.dll mkl_vml_avx512_mic.dll mkl_vml_cmpt.dll mkl_vml_def.dll mkl_vml_mc.dll mkl_vml_mc2.dll mkl_vml_mc3.dll

Build PardisoLoader2.dll, which is a wrapper for MKL's PARDISO solver. PTX files do come with these project, but have to be copied to the application folder. If needed, PTXs can be rebuilt for different CUDA version or compute capability. The following CUDA source projects result in PTXs: collision_kernels, cz_kernels, elem_kernels. Build OpenTK, ManagedCUDA, icFlowLibrary, and SimGUI. Alternatively, OpenTK and OpenTK.GLControl can be obtained via NuGet.

Test simulation setups are available from the following location: https://goo.gl/AvZC4u. They should be copied into '_sims' subfolder in the application path. Additional setups for beam bending tests are available at: https://drive.google.com/file/d/1YN2MsQDhCK_P4bmKtK4HCDFSUDnfX88P/view?usp=sharing