Fast algorithm for the simulation of Ising models

Using a new microcanonical algorithm efficiently vectorized on a Cray XMP, we reach a simulation speed of 1.5 nsec per update of one spin, three times faster than the best previous method known to us. Data for the nonlinear relaxation with conserved energy are presented for the two-dimensional Ising model.     

A new multispin coding algorithm for Monte Carlo simulation of the Ising model

We present a new algorithm for Monte Carlo simulation of the Ising model. The usual serial architecture of a computer is exploited in a novel way, enabling parallel but independent calculations to be carried out on as many spins as there are bits in a computer word in each fundamental move. The algorithm enjoys a substantial increase in execution speed over more usual multispin coding algorithms. By its very nature, the algorithm constitutes a design for a special-purpose processor.     

Fast vectorized algorithm for the Monte Carlo simulation of the random field Ising model

An algorithm for the simulation of the 3-dimensional random field Ising model with a binary distribution of the random fields is presented. It uses multi-spin coding and simulates 64 physically different systems simultaneously. On one processor of a Cray YMP it reaches a speed of 184 million spin updates per second. For smaller field strength we present a version of the algorithm that can perform 242 million spin updates per second on the same machine.     

Compressed quantum simulation of the Ising model

Jozsa et al. [Proc. R. Soc. A 466, 809 2009)] have shown that a match gate circuit running on n qubits can be compressed to a universal quantum computation on log(n)+3 qubits. Here, we show how this compression can be employed to simulate the Ising interaction of a 1D chain consisting of n qubits using a universal quantum computer running on log(n) qubits. We demonstrate how the adiabatic evolution can be realized on this exponentially smaller system and how the magnetization, which displays a quantum phase transition, can be measured. This shows that the quantum phase transition of very large systems can be observed experimentally with current technology.     

Dynamic critical behavior of the worm algorithm for the Ising model

We study the dynamic critical behavior of the worm algorithm for the two- and three-dimensional Ising models, by Monte Carlo simulation. The autocorrelation functions exhibit an unusual three-time-scale behavior. As a practical matter, the worm algorithm is slightly more efficient than the Swendsen-Wang algorithm for simulating the two-point function of the three-dimensional Ising model.     

Monte Carlo simulation of three-dimensional dilute Ising model

A multispin coding program for site-diluted Ising models on large simple cubic lattices is described in detail. The spontaneous magnetization is computed as a function of temperature, and the critical temperature as a function of concentration is found to agree well with the data of Marro et al.⁽⁴⁾ and Landau⁽³⁾ for smaller systems.The first successful epsilon expansion seems to be by D. E. Khmelnitskii,ZhETF 68:1960 (1975), English translationSov. Phys. JETP 41:981 (1975); for numerical estimates see K. E. Newman and E. K. Riedel,Phys. Rev. H25:264 (1982), for experiments see R. J. Birgenau, R. A. Cowley, G. Shirane and H. Yoshizawa,J. Stat. Phys. 34:817 (1984).     

A fast algorithm for the Cyber 205 to simulate the 3D Ising model

We describe a computer program that performs the Metropolis algorithm for the 3D Ising model at a peak speed of 98 million spin updates per second on a 2-pipe CDC Cyber 205. This speed is achieved using the special vector capabilities of the Cyber 205 and multispin coding techniques.     

Monte Carlo simulation of a kinetic Ising model of the glass transition

Results of Monte Carlo experiments for the two-spin facilitated kinetic Ising model on a cubic lattice are presented and compared with a theoretical prediction.     

二维伊辛模型相变临界点温度的模拟计算

用计算模拟方法计算了二维伊辛模型的相变临界点温度,其结果接近严格解,明显布喇格－威廉斯近似和贝特近似的结果。     

Bounds for the anisotropic Ising model

Rigorous upper bounds are found for the magnetisation, susceptibility, critical temperature and crossover exponent in an anisotropic Ising system.     

Monte Carlo simulation of cluster kinetics in three-dimensional Ising model

Simulations in lattices of size 100³ at the critical point of the three-dimensional Glauber kinetic Ising model indicate that the 1935 Becker-Doring equation no longer works there: The growth rates decay in time. These conclusions confirm those in two dimensions.     

Monte Carlo simulation of Ising model on decagonal covering structure

We study the Ising model on a two-dimensional quasilattice developed from the decagonal covering structure. The periodic boundary conditions are applied to a patch of rhombus-like covering pattern. By means of the Monte Carlo simulation and the finite-size scaling analysis the critical temperature is estimated as 2.317±0.002. Two critical exponents are obtained being 1/v=0.992±0.003 and η=0.247±0.002, which are close to the values of the two-dimensional regular lattices as well as the Penrose tilings.     

基于嵌入式GPU的运动目标分割算法并行优化

在光电监视系统中,广泛应用于运动目标分割的PBAS(pixel base adaptive segmenter)算法计算复杂、参数量大,难以达到实时分割的要求。针对PBAS算法是对图像中每个像素点进行独立处理,特别适合于GPU并行加速的特点,对其在嵌入式GPU平台Jetson TX2上进行了并行优化实现。在数据存储结构、共享内存使用、随机数产生机制3个方面对该算法进行了优化设计。实验结果表明,对于480×320像素分辨率的中波红外视频序列,该并行优化方法可以达到132 fps的处理速度,满足了实时处理的要求。     

GPU-based DPSO algorithm for structural optimization of Pt-Co bimetallic nanoparticles

Alloy nanoparticles (NPs) are potential candidates for catalysts in fuel cells applications, and their physical properties are associated with the corresponding stable structures. In this work, a GPU-based discrete particle swarm optimization (DPSO) method is designed to investigate the stable structure of Pt-Co alloy NPs. Register and an optimal block design are used to further increase the acceleration ratio. Comparative experiments on CPU and GPU show that the GPU-based DPSO algorithm possesses superior computational performance, and Register has a great influence on the acceleration ratio. Analyses on the stable structures of Pt-Co NPs demonstrate that Co atoms preferentially locate at vertex and edges of surface, yet Pt atoms exhibit a strong surface and sub-surface segregation only for their composition over 65%.     

Monte Carlo simulation of two-dimensional Ising model with logarithmic exchange interaction

Monte Carlo simulation was done to a 2-D Ising model with exchange interaction depending logarithmically on distance. This model is equivalent to a 2-D Coulomb gas.A configuration corresponding to bound pairs of opposite charges appears below T_c, while above T_c pairs are dissolved into free charges.     

Bethe-Peierls approximation for the disordered Ising model

We study the Ising model for an alloy with an arbitrary number of components. We develop an approximation which reduces to that of Bethe and Peierls when the concentration of one of the components is unity. We investigate within this approximation the dependence of the various thermodynamic quantities, in particularT _c, on the composition of the alloy and the magnetic properties of its constituents. Comparison with the only exact calculation available, that of F. T. Leeet al., for a linear chain, shows extremely satisfactory agreement.Research supported by ARO (D). It has also benefited from the general support of Materials Science at the University of Chicago by the NSF.