Research & Development of Tensor Engine for optimizing quantum mechanic, quantum chemistry based methods on GPUs
The support of the grant VEKOP-2.1.7-15-2016-00163 is gratefully acknowledged.
About the project
Methods in the field of quantum mechanics, its applied fields as well, can be described with tensor contractions. However in a lot of cases the full descriptive power of the tensor contractions can not be fully utilized especially so on GPU hardwares. The aim of the project is to create an engin capable of harnessing the full capacity of tensor expressions and the full computational power of GPUs.
Innovative experimental device and GPU based simulation software for research and development of nano layer based material production
The support of the grant VEKOP-2.1.1-15-2016-00114 is gratefully acknowledged.
About the project
Within the framework of the grant, research and development are being carried out in the field of simulation-assisted material design aimed at producing nano-layer-based materials.
In the 20th century, the chemical industry heavily relied on experimental methods in the research of new materials. Consequently, the production of a material and the optimization of its properties were labor-intensive processes that required a multitude of experiments. Due to the complexity of nanoscale materials used today, we have reached the limits of material development based solely on experiments. However, the remarkable advancement of computational chemistry methods has laid the foundation for the application of simulations. Because of these developments, it has become necessary to support materials science research with modern methods, such as computer-aided design (CAD). The primary tool for this is simulation, which, when used in conjunction with experiments and often as a replacement, allows for the development of materials at significantly lower costs and with a complete understanding of microscopic (atomic and electronic) details.
The project will implement GPU-optimized modules for industrial applications, specifically for the efficient simulation of nanostructures and extended systems (such as crystals, surfaces, micelles) using the so-called ReaxFF empirical reactive force field-based computational chemistry methods. Additionally, a custom-designed laboratory suitable for materials production will be completed. The novelty of the developed implementation lies in the unified application of efficient simulation and flexible manufacturing technologies within a single system.