We have been carried out the large-scale simulation of new functions and novel properties of nano-materials using Earth Simulator. The large-scale simulation has revealed so far the new functions and novel properties of nano-materials through extending the time and space of simulation and increasing the number of atom. This shows that large-scale simulation with high-end computers could be one of effective methodologies for the cutting edge of science and technology. In this research we have also assessed the computational capability of Earth Simulator through applying it to finding novel functions or properties of high temperature superconductors and carbons in nano-scale. It is very clear that Earth Simulator is an effective and powerful computational facility for the large-scale simulation to explore new functions and novel properties of nano-materials.

View full abstract

The prediction of the three-dimensional structures of proteins with the input of only the amino-acid sequence information (prediction from the first principles) is believed to be impossible with the presently available computation power. We present the results of the protein folding simulations on the Earth Simulator. Our simulations are based on generalized-ensemble algorithms, in particular, the replica-exchange molecular dynamics (REMD). We chose a small protein with 56 amino acids (B1 domain of streptococcal protein G). We performed a REMD simulation of protein G in a sphere of water of radius 50 angstroms. The total number of water molecules was 17,187 (the total number of atoms was 52,416 including the protein atoms). Using 112 nodes of the Earth Simulator, we performed a REMD simulation of this system with 224 replicas. The REMD simulation was successful in the sense that we observed a random walk in potential energy space that suggests that a wide conformational space was sampled. This is the largest system that the REMD simulation was successfully performed. We observed the formation of native-like secondary structures (α-helix and β-strands).

View full abstract

A new model for Marangoni convection is presented. This model is a two dimensional (top-view) model which can reproduce a hexagonal pattern which is typical in Marangoni convection. Moreover, in some parameter regime, we can observe fascinating dynamic patterns such as spatiotemporal chaos and traveling hexagons.

View full abstract

A new concept called a boat-sail distance is introduced on the surface of water with flow. That concept is a mathematical model of the shortest time for a boat to move from one point to another against the flow of water. The problem of computing this distance is reduced to a boundary value problem of a partial differential equation, and a numerical method for solving this problem is constructed. The method is a modification of a so-called fast marching method originally proposed for the eikonal equation. Computational experiments show the efficiency and the stableness of the proposal method.

View full abstract