Karamelo is a parallel simulator for the Material Point Method (MPM). It supports both multiple CPUs (master branch) and single GPUs (kokkos branch).
If you would like to stay updated about Karamelo and also help me understand who uses Karamelo, and for what purposes, please fill up this form: https://forms.gle/aHZ9s5Pc4aTQLZ2b8
- Mar 2024: Karamelo now supports GPUs through the use of Kokkos. Please use the dedicated branch kokkos if you want to use it.
- April 2021: Karamelo now supports restarts
- October 2019: It is now possible to model contacts with TLMPM: https://www.researchgate.net/publication/344552510_Modelling_contacts_with_a_total_Lagrangian_material_point_method
- October 2019: Karamelo has been used extensively for the simulations featured in our MPM review: https://www.researchgate.net/publication/343096803_Material_point_method_after_25_years_theory_implementation_and_applications
- January 2020: Karamelo featured in our recent publication on Total Lagrangian MPM: https://www.sciencedirect.com/science/article/pii/S0045782519306759
- April 2020: Karamelo becomes public under the GNU General Public License
- Standard MPM
- Total Lagrangian MPM (https://www.sciencedirect.com/science/article/pii/S0045782519306759)
- CPDI
- Linear
- Cubic Bsplines
- Bernstein quadratic (for total Lagrangian MPM only)
When you publish data obtained with the help of Karamelo, please cite: A. de Vaucorbeil, V. P. Nguyen, and C. Nguyen-Thanh, “Karamelo: an open source parallel c++ package for the material point method,” Computational Particle Mechanics, oct 2020.
This paper is availble only at https://www.researchgate.net/publication/343096803_Material_point_method_after_25_years_theory_implementation_and_applications or on the cite of the editor: https://link.springer.com/article/10.1007/s40571-020-00369-8
Please find Karamelo's documentation here: https://adevaucorbeil.github.io/karamelo/html/index.html