A Neighborhood Grid Data Structure for Massive 3D Crowd Simulation on GPU

Abstract. Nowadays mobile phones, especially smartphones, are equipped with advanced computing capabilities. Most of these devices have multicore processors such as dual-core CPUs and many-core GPUs. These processors are designed for both low power consumption and high performance computation. Moreover, most devices still lack libraries for generic multicore computing usage, such as CUDA or OpenCL. However, computing certain kind of tasks in these mobile GPUs, and other available multicores processors, may be faster and much more efficient than their single threaded CPU counterparts. This advantage can be used in game development to optimize some aspects of a game loop and also include new features. This work presents an architecture designed to process most of the game loop inside a mobile GPU using the Android Renderscript API. As an illustrated test case for the proposed architecture, this work presents a game prototype called "MobileWars", completely developed using the proposed architecture.

show abstract

“…For this work a odd-even sort data structure is used. More on this data structure can be seen on [29,30].…”

Section: The Ngridmentioning

confidence: 99%

MobileWars: A Mobile GPGPU Game

Joselli

Silva²,

Clua³

et al. 2013

Entertainment Computing – ICEC 2013

Self Cite

View full text Add to dashboard Cite

show abstract

“…The framework implements both approaches, a uniform grid based on the works [15], [18], and a neighborhood grid [7], [19]. Figure 3 illustrates the approaches.…”

Section: A the Broad Phasementioning

confidence: 99%

Techniques for designing GPGPU games

Joselli

Silva

Zamith

et al. 2012

2012 IEEE International Games Innovation Conference

Self Cite

View full text Add to dashboard Cite

The increasing level of realism in digital games depends not only on the enhancement of modeling and rendering effects, but also on the improvement of different aspects such as animation, characters artificial intelligence and physics simulation. Normally, games process most of their tasks in the CPU, using the GPU only for graphics processing. Several games and previous works also use GPU computing to process some selected non-graphics subtasks (normally the game physics), while the remaining tasks remain on the CPU. Such a processing partition is frequently pointed to as the bottleneck in these games, as it may induce several expensive data transfers between the CPU and GPU. This paper shows some game examples that execute all their methods entirely on the GPU using of a new GPU computing architecture for keeping the GPU-CPU communication to a minimum. Using the techniques explained in this paper, we present a game that can have an average optimization of about 25% when compared to traditional GPGPU game programing (i.e. only using GPGPU for physics and processing the AI in the CPU) and 200% when compared to traditional CPU programming.

show abstract

“…Recently, many researchers focused on using the GPU power for crowd behavior simulation (in addition to rendering). The reason to that is one can easily simulate a large crowd that a CPU can never achieve in real time . However, GPU‐based crowd simulation techniques utilize almost the full power of the GPU, thus not enough computing resources left for rendering tasks.…”

Section: Literature Backgroundmentioning

confidence: 99%

Hardware‐accelerated dynamic clustering of virtual crowd members

Haciömeroğlu

Ozcan

Barut

et al. 2012

Computer Animation & Virtual

View full text Add to dashboard Cite

In this study, a hardware‐accelerated dynamic clustering of moving virtual entities technique is proposed. By clustering virtual entities, both clustered and unclustered virtual agents became more aware of other agents’ topological configurations. Clustering is based on their continuously changing velocity and position vectors. The proposed clustering technique efficiently uses graphics processor's parallel processing capabilities. Therefore, almost no additional central processing unit overhead is required to bring the clustering information into the simulation. In addition, in this paper, how cluster information can be used on top of the proposed virtual human steering technique is explained. The results show that by using the dynamic clustering, the number of collision in the simulation reduces, and the velocities of the the agents in the simulation increase. Copyright © 2012 John Wiley & Sons, Ltd.

show abstract

A Neighborhood Grid Data Structure for Massive 3D Crowd Simulation on GPU

Cited by 21 publications

References 11 publications

MobileWars: A Mobile GPGPU Game

MobileWars: A Mobile GPGPU Game

Techniques for designing GPGPU games

Hardware‐accelerated dynamic clustering of virtual crowd members

Contact Info

Product

Resources

About