CUDA Software Research Articles

Thermodynamic equilibrium of [formula omitted] Ising model on square lattice

We constructed a theoretical magnetic phase diagram in an external magnetic field T(P+), making it possible to determine the conditions for the existence of ferromagnets, antiferromagnets, and spin glass phases. The high-performance CUDA software package was used to the complete enumeration of all configurations of finite number spins in the ±J Ising model. We performed the rigorous numerical calculation of the partition function of N=8×8 systems of interacting spins with open boundary conditions. We used Monte Carlo methods like the Metropolis algorithm to calculate the critical temperatures for N=40×40 spins. The results of the Monte Carlo experiments are consistent with rigorous calculation data. The transition from the spin glass to the induced ferromagnetic state in an external field occurs without any critical change in the heat capacity. We used the ±J Ising model to calculate the instability line (AT—line) for the heat capacity of spin glass in the H−T diagram in an external magnetic field and the behavior of magnetic susceptibility in an external magnetic field. A rigorous calculation of the partition function allowed us to calculate all possible states and their thermodynamic probability. The calculation of the partition function meant that the model’s physics was obtained in an equilibrium state. The instability line was calculated for spin glass in the equilibrium state.

РЕАЛИЗАЦИЯ ПАРАЛЛЕЛЬНЫХ ВЫЧИСЛЕНИЙ НА ОСНОВЕ ПРОГРАММНО-АППАРАТНОГО СТЕКА CUDA

This paper describes in detail the problems of implementing parallel computing based on the CUDA hardware and software stack. It provides a comprehensive analysis of the key aspects of using CUDA for parallel programming and data processing. The paper examines in detail the structure of the CUDA architecture, describes the programming model and the optimization process

OPTIMIZATION OF DESIGN OF THE FISH-SPAWNING PASS OF THE BAGAEVSKY HYDROELECTRIC COMPLEX USING NUMERICAL 2D MODELING

Numerical hydrodynamic modeling of the flow of water and sediment through the fish-spawning pass of the Bagaevsky hydroelectric complex on the river Don is considered. The calculations use two-dimensional shallow water and sediment transport equations by the STREAM 2D CUDA software complex. Recommendations are made to change the longitudinal profile of the bottom and the configuration of the inlet section, as well as to increase the coefficient of roughness of the bottom and sides of the channel to ensure fish protection requirements.

NUMERICAL STUDY OF ICE AND CHANNEL PROCESSES ON THE BRIDGE SUPPORTS ACROSS THE KUIBYSHEV RESERVOIR (KAZAN)

Numerical hydrodynamic modeling of water movement at discharges of 1% and 10% supply is considered in order to study the effect of ice effects and bottom deformations on the bridge supports across the reservoir of the Kuibyshev hydroelectric complex, and to predict the general erosion of the channel. The calculations use two-dimensional shallow water and sediment transport equations implemented in the STREAM 2D CUDA software package.

Assessing Inundation Risks and Protection of Residential Territories in River Floodplains

It is shown that the likely damage caused by inundation of residential areas and objects of higher hazard on river floodplains can be evaluated taking into account the inundation depth. A procedure of constructing inundation probability curves with the use of two-dimensional hydrodynamic models (based on the STREAM 2D CUDA software system (Russia)) and field observation data is considered. Data from hydrological gages in several rivers are used to show that the inundation recurrence curves can be approximated by some logarithmic functions. Formulas for determining the inundation risks of residential areas and integral risk characteristics are derived and substantiated. It is shown that, on the average for the objects considered in the study, the maximal risks correspond to inundation levels with exceedance probability of ~2%, and that the effective protection of residential areas in most cases should by designed for inundation levels with much smaller exceedance probability.

Моделирование транспорта наносов и русловых переформирований в районе Сормовского водозабора в условиях половодий, межени и при работе земснарядов

The construction and operation of hydraulic structures has an impact on the environment and nearby objects. The article discusses the Sormovo intake structure and the possible impact on it of the projected Nizhny Novgorod low-pressure hydroelectric complex. The STREAM 2D CUDA software package, based on the original numerical algorithm for solving two-dimensional shallow water equations (Saint-Venant equations) on an uneven bottom, was used as a research tool. To calculate the deformations of the river bottom and siltation of water intake structures, a new physical and mathematical model of sediment transport is used, taking into account the heterogeneous (multi-fractional) composition of the soil, the movement of both traction and suspended sediments, and the reformation of underwater and above-water slopes. Three models were built for the calculations. At the fi rst one, calculations were made for fl oods of rare frequency of 10%, 1%, and 0.1% probability. On the second low water conditions – the navigation period with daily regulation of the above located hydroelectric complex. The third model considered the construction period, when hydro mechanized works are carried out. Based on the calculations obtained, plans for erosion – accumulation, plans for sediment deposition were built, and the masses of sediments entering the Sormovsky intake were determined. The simulation results showed that the bottom deformations in the intake structure area are insignifi cant. Most of the sediment enters the intake structure bucket during the flood period. At the same time, the construction of the Nizhny Novgorod low-pressure hydroelectric complex does not change the existing trend much. The main conclusion of the work done is that special additional measures to protect the Sormovo intake structure from sediments during the navigation period will not be required.

Selection of contour images based on alphabetical representation

The article is devoted to the actual problem of searching for contour images by given reference samples, due to the large amount of graphic scientific and technical information of the Contour type. Existing methods for image search do not allow taking into account the actual identity of images arising from affine transformations (stretching / compression, rotation) resulting from replication, which is important when analyzing graphic documentation. The paper proposes an alphabetical representation of contour images, which made it possible to construct a complex of affine-invariant equivalence relations. The relations built up served as the basis for the development of a set of selection procedures that can be considered as the initial stage of image search. The procedures included in the complex sequentially implement the reduction of the search scope up to a small set of contour images, allowing for the possibility of visual analysis. The complex is programmatically implemented in Python using the CUDA (NVIDIA GPU) software and hardware architecture for parallel computing.

Ice impact on the flow dynamics the Northern Dvina River mouth area

The authors consider case of the winter period 2019-2020 as example of ice cover and ice jam influence on the flow dynamics in the section upstream the delta of the Northern Dvina River mouth area. Hydrological monitoring, calculations and forecasts of the ice drift in the Northern Dvina River mouth area are actual practical tasks. The hydrological regime of a tidal river mouth is very complex phenomena and one of the most difficult for a quantitative description. The authors implemented an integrated approach to study ice jams, including field ice surveys, satellite data analysis, data from the state hydrological monitoring system and hydrodynamic modeling. STREAM_2D CUDA software is applied for two-dimensional numerical modeling. It takes into account the impact of ice cover. Ice phenomena usually make significant changes in the water level and velocity regime. In 2019/2020 hydrological year the ice cover provided and maintained a large backwater during all the autumn and the winter. When the spring ice drift came from the river upstream there was a 3-4 meter water level rise, which lasted for 5 days, creating a threat of flooding of the Kholmogory village by a breakthrough wave. There is no effective technique to forecast or assess probable scenarios for such situation. The article identified difficulties arising during ice conditions analysis and hydrodynamic modeling of the situation. Steps to optimize the monitoring of ice cover, flow discharges and water levels are also formulated

Numerical Modeling of a Hydrodynamic Accident at an Earth-and-Rockfill Dam on the Dyurso River

The main aim of the study was to perform a modeling of a hydrodynamic accident which took place in early August 2002 at an earth-and-rockfill dam near the city of Novorossiysk on the Dyurso River. The study was carried out by applying numerical modeling methods in a two-dimensional plan setting using the STREAM 2D CUDA software package. In this study, the development of a washout affecting an earth dam made of a heterogeneous material was modeled for the first time.

Numerical Modeling of the Behavior of a Destructive Rain Flood on a Mountain River

The objective of this study was to provide the most accurate presentation of the behavior of a disastrous rain flood, which had resulted in destruction of a dam, casualties, and significant material damage. The problems set were solved by methods of numerical modeling in the two-dimensional setting applying the STREAM 2D CUDA software package. To calculate the rain flood, a catchment model of the Durso River was developed on a triangular non-uniform mesh, adapted to the river channel and the main inflows. To ensure direct numerical modeling of the dam’s destruction, a model was developed, which included the water area of the reservoir, the dam, composed of non-homogeneous earth material, and a downstream section of the river from the dam to the river mouth. The main results of the study were: the hydrograph of the water discharge of the rain flood, calculated by the actual precipitation data; a description of the washout of the earth dam composed of non-homogeneous materials, resulting from water spill over the dam crest, and a desctiption of the downstream spread of the breach wave.

METHOD OF SOLVING SYSTEM OF LINEAR ALGEBRAIC EQUATIONS WITH USE OF HIGH PERFORMANCE COMPUTING ON GPU

The article discusses the use of graphics processing units for solving large system of linear algebraic equations (SLAE). A heterogeneous multiprocessor computing platform produced by the NIIVK, whose architecture allows the integration of general‑ purpose microprocessor modules with graphics processor modules was used as an equipment for solving SLAEs. The description of the SLAE solution program, developed on the basis of the CUBLAS CUDA software interface library, is given. A method is proposed for increasing the accuracy of calculations of linear systems based on the use of a modified Gauss method. It has been established that the use of the modified Gauss method practically does not increase the program operation time with a significant increase in the accuracy of calculations. It is concluded that the use of graphics processors for solving SLAEs allows processing matrices of a larger size compared to the use of general‑purpose microprocessors.

Heterogeneous parallel computing accelerated generalized likelihood uncertainty estimation (GLUE) method for fast hydrological model uncertainty analysis purpose

The generalized likelihood uncertainty estimation (GLUE) is a famous and widely used sensitivity and uncertainty analysis method. It provides a new way to solve the “equifinality” problem encountered in the hydrological model parameter estimation. In this research, we focused on the computational efficiency issue of the GLUE method. Inspired by the emerging heterogeneous parallel computing technology, we parallelized the GLUE in algorithmic level and then implemented the parallel GLUE algorithm on a multi-core CPU and many-core GPU hybrid heterogeneous hardware system. The parallel GLUE was implemented using OpenMP and CUDA software ecosystems for multi-core CPU and many-core GPU systems, respectively. Application of the parallel GLUE for the Xinanjiang hydrological model parameter sensitivity analysis proved its much better computational efficiency than the traditional serial computing technology, and the correctness was also verified. The heterogeneous parallel computing accelerated GLUE method has very good application prospects for theoretical analysis and real-world applications.

GPU based parallel genetic algorithm for solving an energy efficient dynamic flexible flow shop scheduling problem

Due to new government legislation, customers’ environmental concerns and continuously rising cost of energy, energy efficiency is becoming an essential parameter of industrial manufacturing processes in recent years. Most efforts considering energy issues in scheduling problems have focused on static scheduling. But in fact, scheduling problems are dynamic in the real world with uncertain new arrival jobs after the execution time. This paper proposes an energy efficient dynamic flexible flow shop scheduling model using the peak power value with consideration of new arrival jobs. As the problem is strongly NP-hard, a priority based hybrid parallel Genetic Algorithm with a predictive reactive complete rescheduling strategy is developed. In order to achieve a speedup to meet the short response in the dynamic environment, the proposed method is designed to be highly consistent with the NVIDIA CUDA software model. Finally, numerical experiments are conducted and show that our approach can not only solve the problem flexibly, but also gain competitive results and reduce time requirements dramatically.

Effective NC machining simulation with OptiX ray tracing engine

The manufacturing of high-added-value products in multi-axis machining requires advanced simulation in order to validate the process. Whereas CAM software editors provide simulation software that allows the detection of global interferences or local gouging, research works have shown that it is possible to consider multi-scale simulations of the surface, with a realistic description of both the tools and the machining path. However, computing capacity remains a problem for interactive and realistic simulations in 5-axis continuous machining. In this context, using general-purpose computing on graphics processing units as well as NVIDIA OptiX ray tracing engine makes it possible to develop a robust simulation application. Thus, the aim of this paper is to evaluate the use of NVIDIA OptiX ray tracing engine compared to a fully integrated CUDA software, in terms of computing time and development effort. Experimental investigations are carried out on different hardware such as Xeon CPU, Quadro4000, Tesla K40 and Titan Z GPUs. Results show that the development of such an application with the OptiX development kit is very simple and that the performances in roughing simulations are very promising. Developed software as well as dataset can be downloaded from http://webserv.lurpa.ens-cachan.fr/simsurf .

Portable Solution for Modeling Compressible Flows on All Existing Hybrid Supercomputers

A variant of a numerical algorithm for simulating viscous gasdynamic flows on unstructured hybrid grids and its software implementation for heterogeneous computations is described. The system of Navier–Stokes equations is approximated by the finite-volume method of an increased approximation order with the values of the variables being defined at the mass centers of the grid elements. The distributed software implementation of the numerical algorithm is adapted to running on hybrid computer systems of various architectures. Comparative implementations were created using the MPI, OpenMP, CUDA, and OpenCL software models permitting the use of multicore processors and various types of accelerators, including NVIDIA and AMD graphics processors, and Intel Xeon Phi multicore coprocessors. The data exchange between MPI processes and between processors and accelerators is carried out simultaneously with the execution of calculations (both in MPI + OpenMP mode and when using CUDA or OpenCL). The indicators of parallel efficiency and performance on systems with different types of computing devices are studied in detail. In the tests, up to 260 GPUs were successfully used.

Parallel algorithms for reducing derivation time of distinguishing experiments for nondeterministic finite state machines

Many approaches have been proposed for deriving tests from finite state machine (FSM) specifications with respect to some established coverage criteria. A fundamental core problem in FSM-based testing relates to the derivation of input sequences that can distinguish states of an FSM specification, aka distinguishing sequences. A major effort in the construction of these sequences is based on the derivation of a successors search-tree labeled by sets of pairs of states of the given machine. We aim at reducing the time associated with such constructions through the use of state-of-the-art parallel technologies. Namely, we propose a parallel algorithm that we implement and evaluate on multicore CPUs and on many-core GPUs. We evaluate two alternative GPU implementations that use the CUDA and Thrust software platforms and a network of workstations based solution. The latter sports a workload partitioning based on Divisible Load Theory. A rigorous set of experiments highlights the differences of the proposed implementations in terms of execution time and speedup.

Estimation of the physico-chemical parameters of materials based on rare earth elements with the application of computational model

Computational model, technique and the basic principles of operation program complex for quantum-chemical calculations of material's physico-chemical parameters with rare earth elements are discussed. The calculating system is scalable and includes CPU and GPU computational resources. Control and operation of computational jobs and also Globus Toolkit 5 software provides the possibility to join computer users in a unified system of data processing with peer-to-peer architecture. CUDA software is used to integrate graphic processors into calculation system.

Buffer overflow vulnerabilities in CUDA: a preliminary analysis

We present a preliminary study of buffer overflow vulnerabilities in CUDA software running on GPUs. We show how an attacker can overrun a buffer to corrupt sensitive data or steer the execution flow by overwriting function pointers, e.g., manipulating the virtual table of a C++ object. In view of a potential mass market diffusion of GPU accelerated software this may be a major concern.

GPU Application in Cuda Memory

Nowadays modern computer GPU (Graphic Processing Unit) became widely used to improve the performance of a computer, which is basically for the GPU graphics calculations, are now used not only for the purposes of calculating the graphics but also for other application. In addition, Graphics Processing Unit (GPU) has high computation and low price. This device can be treat as an array of SIMD processor using CUDA software. This paper talks about GPU application, CUDA memory and efficient CUDA memory using Reduction kernel. High-performance GPU application requires reuse of data inside the streaming multiprocessor (SM). The reason is that onboard global memory is simply not fast enough to meet the needs of all the streaming multiprocessor on the GPU. In addition, CUDA exposes the memory space within the SM and provides configurable caches to give the developer the greatest opportunity of data reuse.

An automated algorithm for lesion identification in dynamic contrast enhanced MRI

Breast cancer is one of the most common cancers found in women around the world. Dynamic contrast enhanced magnetic resonance imaging DCE-MRI is nowadays used in many suspected cases that may not be visible in a mammogram. This work concentrates on algorithmic development for segmentation of lesions, automating region of interest ROI, and to delineate the malignancy on DCE-MRI. This has been tested against the intensity time kinetic curve method. The curve pattern and its distribution shows the likelihood of malignancy, which can be automatically coloured based on the assessment. The system helps to compute volume rendering by forming a 3D structure of the lesion area. The proposed system enables doctors to improve the diagnostic accuracy. The development is carried out using graphic processing unit GPU based high end parallel computing with NVIDIA CUDA software, which enables to process more than 5000 images of a single patient with better speed and managed memory.