Optimizing Memory Access Efficiency in CUDA Kernel via Data Layout Technique

Over the past decade, Graphics Processing Units (GPUs) have revolutionized high-performance computing, playing pivotal roles in advancing fields like IoT, autonomous vehicles, and exascale computing. Despite these advancements, efficiently programming GPUs remains a daunting challenge, often relying on trial-and-error optimization methods. This paper introduces an optimization technique for CUDA programs through a novel Data Layout strategy, aimed at restructuring memory data arrangement to significantly enhance data access locality. Focusing on the dynamic programming algorithm for chained matrix multiplication—a critical operation across various domains including artificial intelligence (AI), high-performance computing (HPC), and the Internet of Things (IoT)—this technique facilitates more localized access. We specifically illustrate the importance of efficient matrix multiplication in these areas, underscoring the technique’s broader applicability and its potential to address some of the most pressing computational challenges in GPU-accelerated applications. Our findings reveal a remarkable reduction in memory consumption and a substantial 50% decrease in execution time for CUDA programs utilizing this technique, thereby setting a new benchmark for optimization in GPU computing.

Optimization of horizontal well staggered patterns

Staggered line-drive patterns are widely used in oilfields. In this paper, to optimize a staggered pattern of horizontal wells, a 3D problem was divided into two 2D （x-y plane andy-z plane） problems with the pseudo-3D method, conformal transformation and superposition principle. A productivity equation for a horizontal well was deduced, which can be used to optimize the well pattern. A relationship between the length of horizontal wells and the shape factor of well patterns was established. The result shows that optimized well patterns can improve oil production from horizontal wells. This provides a theoretical basis for horizontal well applications to the development of oilfields, especially for overall development of oilfields by horizontal wells.

Ecologically based landscape pattern optimization in northwest of Beijing

Supported by the technologies of remote sensing(RS) and geographical informa-tion system(GIS),we chose northwest of Beijing as a study area and gave priority to under-standing of the spatial-temporal characteristics of landscape pattern change through visually interpreted Landsat TM images of 1989,1996 and 2005.It is believed that there were a series of landscape ecological problems caused by city expansion:landscape ecological connec-tivity was low;landscape structure was simplified;the fragmentation of green land patch was more obvious on the plain areas,moreover,spatial distribution of green land was unbalanced.For this reason,this study adopted accumulative cost distance model,combined with eco-system services and spatial interactions of landscape types,analyzed the spatial difference of the ecological function and the compactness of landscape structure in the study area,and further discussed the landscape pattern optimization proposal.We find that it is essential to protect and establish ecological sources,to establish urban ecological corridors,and to es-tablish ecological nodes at the landscape ecological strategic positions so as to intensify spatial relationships among landscape elements and maintain continuity of landscape eco-logical process and pattern in the course of city expansion.The methods and final results from this study are expected to be useful for landscape ecological planning in Beijing region.

The ICM method with objective function transformed by variable discrete condition for continuum structure

ICM （Independent Continuous Mapping） method can solve topological optimization problems with the minimized weight as the objective and subjected to displacement constraints. To get a clearer topological configuration, by introducing the discrete condition of topological variables and integrating with the original objective, an optimal model with multi-objectives is formulated to make the topological variables approach 0 or 1 as near as possible, and the model reduces the effect of deleting rate on the result. The image-filtering method is employed to eliminate the checkerboard patterns and mesh dependence that occurred in the topology optimization of a continuum structure. The computational efficiency is enhanced through selecting quasi-active displacement constraints and a design region. Numerical examples indicate that this algorithm is robust and practicable, though the number of iterations is slightly increased with respect to the original algorithm.

Adaptive backtracking search optimization algorithm with pattern search for numerical optimization

The backtracking search optimization algorithm（BSA） is one of the most recently proposed population-based evolutionary algorithms for global optimization. Due to its memory ability and simple structure, BSA has powerful capability to find global optimal solutions. However, the algorithm is still insufficient in balancing the exploration and the exploitation. Therefore, an improved adaptive backtracking search optimization algorithm combined with modified Hooke-Jeeves pattern search is proposed for numerical global optimization. It has two main parts： the BSA is used for the exploration phase and the modified pattern search method completes the exploitation phase. In particular, a simple but effective strategy of adapting one of BSA＇s important control parameters is introduced. The proposed algorithm is compared with standard BSA, three state-of-the-art evolutionary algorithms and three superior algorithms in IEEE Congress on Evolutionary Computation 2014（IEEE CEC2014） over six widely-used benchmarks and 22 real-parameter single objective numerical optimization benchmarks in IEEE CEC2014. The results of experiment and statistical analysis demonstrate the effectiveness and efficiency of the proposed algorithm.

Landscape Pattern Optimization of Chengdu Based on Modern Garden City

TOptimization of regional landscape pattern is significant for improving function and value of ecosystem,and restraining the expansion of urban layout.Taking Chengdu City for example,this paper applied RS and GIS techniques,landscape indexes and ecological service function evaluation to further analyze the temporal and spatial characteristics of landscape pattern and spatial differences of regional ecological functions,and on this basis,identified the spatial distribution of ecological source lands.Based on the long-term objective of building Chengdu into a modem garden city,this paper applied the accumulative cost distance model and introduced garden city theory to construct regional ecological corridors and ecological nodes,and explored the approaches of optimizing landscape pattern of modem garden city.The results showed that a great deal of arable land has been transferred to construction land in the urbanization;intensity of regional ecological functions showed obvious spatial differences;ecological source lands were mainly distributed in the Longmen Mountain,the Qionglai Mountain,the Changqiu Mountain and the Longquan Mountain;according to actual conditions of the study area,the road ecological corridors,river corridors and agricultural corridors in the layout of "four rings and six radial corridors" were constructed;ecological nodes dominated by intersection,wetland and forest park were formed.This research method and results are significant references for building Chengdu into a modem garden

Novel Optimization Approach to Mixing Process Intensification

An approach was presented to intensify the mixing process. Firstly, a novel concept, the dissipation of mass transfer ability(DMA) associated with convective mass transfer, was defined via an analogy to the heat-work conversion. Accordingly, the focus on mass transfer enhancement can be shifted to seek the extremum of the DMA of the system. To this end, an optimization principle was proposed. A mathematical model was then developed to formulate the optimization into a variational problem. Subsequently, the intensification of the mixing process for a gas mixture in a micro-tube was provided to demonstrate the proposed principle. In the demonstration example, an optimized velocity field was obtained in which the mixing ability was improved, i.e., the mixing process should be intensified by adjusting the velocity field in related equipment. Therefore, a specific procedure was provided to produce a mixer with geometric irregularities associated with an ideal velocity.

Study on Optimized Pattern of Forest Ecosystem Management in Dagangshan

A comprehensive analysis was made based on the relevant information from the results of the orientation study by the Dagangshan Forest Ecosystem Research Station obtained in the past 20 years The conclusions reached from the study are as follows: The biodiversity in the area is constantly on the decline and the categories of forest, forest age and tree species were out of balance. The stumpage had reduced by 18.1% from 77.4 m3/ha to 63.4 m3/ha in this period. The ecological benefits declined and the com...

Evaluation of infill well pattern based on the dynamic change of reservoirs during coalbed methane development

With the deepening of coalbed methane(CBM)exploration and development,the problem of low gas production has gradually become one of the main factors restricting the development of the CBM industry in China.Reasonable well pattern deployment can improve the productivity of CBM wells and reduce the cost of production,while the reservoir changes of CBM wells play a important role for well pattern infilling.In this study,the dynamic characteristics of the average reservoir pressure(ARP),permeability,and drainage radius during the development process of CBM wells are systematically analyzed,and predicted the production changes of well groups before and after infilling wells in combination with the characteristics of reservoir changes.The results show that the high gas production wells have a larger pressure drop,long drainage radius,and a large increase in permeability.On the contrary,low gas production wells are characterized by small drainage radius,damaged permeability and difficult to recover.The productivity of infilled horizontal wells is predicted for two well groups with different productivity and reservoir dynamic characteristics.After infilling wells,the production of current wells has increased at different degrees.It is predicted that the average gas production of low gas production well group H1 and middle gas production well group H2 will increase 1.64 and 2.09 times respectively after 3000 days production simulation.In addition,the pressure interference between wells has increased significantly,and the overall gas production of the well group has greatly increased.Infill wells can achieve better development results in areas with superior CBM resources,recoverable reservoir permeability,and small drainage radius during the early production process.The research results provide a reference for the later infill adjustment of CBM well patterns in the study area.

COMPLEXITY ANALYSIS AND COMPUTATION OF THE OPTIMAL HARVESTING FOR ONE-SPECIES POPULATION RESOURCES

The exploitation of renewable resources creates many complex problems for culture, ecology and economics as well. Ascertaining the essentials behind the coraplex problems is very important, in this paper, we mainly study various complex relations appearing in the optimal exploitation process for renewable resources. First, we derive a sufficient condition on the existence of optimal harvesting policies for one-species population resources. Then we present every possible optimal harvesting pattern for such a modeh On the basis of this, we give a computing formula for estimating the optimal harvesting period, optimal transitional period, and optimal recruitment period. The main difference with respect to the previous works in literature is that our optimal harvesting policy is a piece-wise continuous function of time t, at the piecewise point to, which is called switching time. At the switching time we switch the harvesting rate from h to some transitional control u,, then to 0. Clearly this kind of harvesting policy is easier to carry out than those by others, provided that there exists a managing department which can highly supervise the resources.

Cropping pattern optimization considering uncertainty of water availability and water saving potential

In arid and semi-arid areas,the profitability of irrigated agriculture mainly depends on the availability of water resources and optimal cropping patterns of irrigation districts.In this study,an integrated agricultural cropping pattern optimization model was developed with considering the uncertainty of water availability and water saving potential in the future,aiming to maximize agricultural net benefit per unit of irrigation water.The available water which was based on the uncertainty of runoff was divided into five scenarios.The irrigation water-saving potential in the future was quantified by assuming an increase in the rate irrigation water-saving of 10% and 20%.The model was applied to the middle reaches of Heihe River basin,in Gansu Province,China.Results showed that if the irrigation water-saving rate was assumed to increase by 10%,then the net water-saving quantity would increase by 21.5-22.5 million m3 and the gross water-saving quantity would increase by 275.7-303.0 million m3.Similarly,if the irrigation water-saving rate increased by 20%,then the net water-saving quantity would increase by 43.0-45.1 million m3 and the gross water-saving quantity would increase by 331.7-383.2 million m3.If the agricultural cropping pattern was optimized,the optimal water and cultivated area allocation for maize would be greater than those for other crops.Under the premise that similar volume of irrigation water quantity was available in different scenarios,results showed differences in system benefit and net benefit per unit of irrigation water,for the distribution of available irrigation water was diverse in different irrigation districts.

Analysis of simultaneous effect of water access requirements in Regional Water Corporation ofWest- Azerbaijan and risk factor on cultivation optimal pattern

In vast farms management,especially in multi-crop cultivation case,planning for a combination of crops cultivation has always been of high importance to achieve the maximum income,by taking into account imposed limitations and the risk of different activities.According to the fact that in traditional models of farm programming because of not attention to risk,its results differ from what farmer do in reality,considering the risk farm programming should be well thought.MOTAD model(Minimization of the Total Absolute Deviation)is a mathematic programming model that enters risk into decision making related to farm activities and provides different programs.A main feature of this model is that the risk is calculated by negative deviation from efficiency amount as total multiplication.In recent years,serious declining of water table problem forced Regional Water Corporation of West-Azerbaijan province to impose limitations for extra water use of wells in order to sustain"Mako,Poldasht"water resources.These limitations result in lowering available water level(The most important input for farming),and create changes in cropping patterns and income level of farmers in this area.The aim of this study is analyzing improved cropping considering water limitations and in risk conditions.Data have been collected from farmers and Agriculture"Jihad"Management of town of Mako for duration of 1387-1388.According to the results of the research,the real cultivation plan that is now being implemented throughout the region carries a high risk and low income.Based on results,if the farmer uses the optimal cultivation pattern during the so-called water access limitation,watermelon harvest will be excluded from the optimal cultivation pattern.Also,alfalfa will enter the cultivation plan with the expected income level of 1067 billion Rials(100 Rials=US$0.01).If the farmer uses improved cropping pattern in limitations in risk it will decrease total income as 171.36 billion Rials,to when the water limitations is not implemented.

Ecosystem Service Value Evolution and Security Pattern Optimization in Huaihai Economic Zone

The ecological and environmental effects caused by land use change have attracted global attention.Huaihai Economic Zone, as the core of the Huaihe River ecological economic belt, has experienced a reciprocal evolution of land use, ecological security and regional economic development. Based on multi-stage land use data extracted by Google Earth Engine(GEE), the spatio-temporal differentiation characteristics of ecosystem service value(ESV) evolution in Huaihai Economic Zone from 1998 to 2018 were analyzed with the help of ESV assessment and a minimum accumulated resistance model(MCR), and the regional ecological security pattern(ESP) was optimized. The results show that ESV intensity has obvious spatial differentiation, which is higher in northeastern China and lower in southwestern China. The median ESV area accounted for the largest proportion, while the high and low ESV areas accounted for a small proportion. The characteristics of EVS temporal and spatial differentiation show decreasing and increasing grades. From the perspective of development period, the ESV grade changes show a positive trend. In the optimization of the ecological security pattern, 26 important ecological sources, 22main landscape ecological corridors, and 65 ecological strategic nodes were optimized and identified, and the middle-level ecological security zone accounted for the largest proportion. The main reasons for the changes in the ESV and ESP are closely related to the changes in local natural resources and the changes and adjustments in government protection policies. These research results can provide a reference for inter-provincial territorial space protection and the formulation of a sustainable development strategy.

Task-oriented Hierarchical Control of Modular Soft Robots with External Vision Guidance

General,high-precision theoretical modeling method is not well developed in the field of soft robotics,which holds back motion control and practical application of soft robots.The concept of modularization brings novel structure,novel locomotion patterns as well as novel control method for soft robots.This paper presents the concept of hierarchical control method for modular soft robot system and a H-configuration pneumatic modular soft robot is designed as the control object.The H-configuration modular soft robot is composed of two basic motion units that take worm-like locomotion principle.The locomotion principle of the basic motion unit is analyzed and the actuation sequence is optimized by evolution strategy in VOXCAD simulation software.The differential drive method is applied to the H-configuration modular soft robot with multi motion modes and vision sensor is used to control the motion mode of the robot.The H-configuration modular soft robot and the basic motion unit are assembled by a cubic soft module made of silicone rubber.Also,connection mechanism is designed to ensure that the soft modules can be assembled in any direction and posture.Experiments are conducted to verify the effect of the hierarchical control method of the modular soft robots.