• Electronics and Telecommunications Trends
• /
• v.31 no.4
• /
• pp.65-76
• /
• 2016

특정 제조사의 단말 또는 운영체제에 의존성이 없는 플랫폼 독립적인 웹은 높은 이식성, 소프트웨어의 재활용, 개발 생산성, 풍부한 개발자 존재, 유지 보수 면에서 장점을 가지나, 화려한 UI/UX를 제공하는 네이티브 응용에 비해 낮은 성능으로 웹 기반의 응용 개발 및 보급이 크게 활성화되지 못했다. 한편 데스크톱은 물론 모바일 단말의 멀티코어 기반 Graphic Processing Unit(GPU), CPU 탑재 등 HW 고사양화와 웹 응용에서도 HW 가속 기능을 활용할 수 있는 표준 제공으로 성능 제약을 극복할 수 있게 되었다. 본고에서는 GPU 발전동향을 살펴보고, 고속 렌더링 및 병렬 연산처리를 요구하는 웹 응용이 GPU기반 HW 가속 기능을 활용할 수 있는 크로노스 그룹의 그래픽 가속(Web Graphics Library: WebGL) 및 컴퓨팅(Web Computing Language: WebCL) 지원 표준 규격을 정리한다. 또한, 최근 차세대 GPU Application Programming Interface(API)로 발표된 Vulkan에 대해 알아보고, 웹 고속화 기술에 적용 가능성에 대해 전망한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06a
• /
• pp.118-120
• /
• 2006

최신 GPU는 일반 CPU보다 10배 이상 빠른 연산능력을 갖추고 있는데다가 사용자가 직접 프로그래밍 할 수 있기 때문에 이를 이용한 고속 볼륨 렌더링 알고리즘에 대한 연구가 활발하게 진행되고 있다. 그러나 스트림 프로세싱에 특화 돼있는 GPU의 특성상 early ray termination과 empty space skipping을 구현하는 것이 쉽지만은 않다. 특히 지금까지 제안됐던, 프록시 도형(proxy geometry)을 사용하는 볼륨 렌더링 알고리즘은 empty space skipping은 비교적 효율적으로 구현하지만 early ray termination의 지원은 상대적으로 미비했다. 본 논문에서는 스텐실 버퍼와 OpenGL 확장(extension)을 이용한 2-Pass 알고리즘을 통해서 early ray termination과 empty space skipping을 동시에 구현하는 방법을 제시하고, 그 성능을 측정했다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06c
• /
• pp.37-40
• /
• 2011

계층적 색인 구조는 대용량의 다차원 데이터에 대한 범위질의를 가장 효율적으로 처리하는 색인 구조이다. 계층적 색인 구조에서 범위질의의 속도를 향상시키기 위해서 색인 구조의 구성 시 발생하는 인접노드간의 겹치는 영역을 줄이는 기법들과 다량의 데이터를 한 번에 읽어 상향식 방식으로 색인 구조의 공간 활용도를 증가시키는 벌크 로딩 기법들이 제안되었다. 하지만 CPU기반에서 개별의 노드들을 순차적으로 질의처리 하는 계층적 색인 구조는 공간 활용도의 증가와 노드 간의 중첩 영역을 줄이는 것만으로는 질의 처리 성능 향상에 한계가 있다. 따라서 본 논문에서는 기존의 CPU기반 계층적 색인 구조 중의 대표적인 예인 R-tree의 저장 구조를 GPU 메모리에 적합하도록 변경을 하였다. 또한 기존 CPU기반 계층적 색인 구조의 순차적인 노드 검색을 GPU를 이용해 병렬적으로 노드를 검사하여 성능을 향상시켰다. 이와 같은 방식으로 질의 영역의 크기에 따라서 성능 향상정도가 다르지만 최대 100배 이상의 성능을 향상시켰다.

• 한국HCI학회:학술대회논문집
• /
• 2009.02a
• /
• pp.455-458
• /
• 2009

The statistical wisp model for hairstyle generation was introduced in [1]. It provided a program to load human models, set parameters, generate wisps and strands, and make constraints. However, the program used hard-coded human models and prescribed constraints so that it was hard to change different models and manipulate constraints. Hence we provide a simple interface by drawing maps and constraints. Also, we can increase the speed of computation by using GPU acceleration.

• Journal of Korea Multimedia Society
• /
• v.17 no.1
• /
• pp.69-76
• /
• 2014

It is hard to represent massive terrain data in real-time even using recent graphics hardware. In order to process massive terrain data, mesh simplification method such as continuous Level-of-Detail is commonly used. However, existing GPU-based methods using quad-tree structure such as geometry splitting, produce lots of vertices to traverse the quad-tree and retransmit those vertices back to the GPU in each tree traversal. Also they have disadvantage of increase of tree size since they construct the tree structure using texture. To solve the problem, we proposed GPU-base chunked LOD technique for real-time terrain rendering. We restrict depth of tree search and generate chunks with tessellator in GPU. By using our method, we can efficiently render the terrain by generating the chunks on GPU and reduce the computing time for tree traversal.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.49 no.8
• /
• pp.1-9
• /
• 2012

Quantum-Inspired Evolutionary Algorithm(QEA) contains sufficient data-level parallelism to be naturally accelerated on GPUs. For an efficient reduction of execution time, however, careful task-mapping should be done to properly reflect the characteristics of CPU and GPU. Furthermore, when deciding which part of the application should run on GPU, we need to consider the data transfer between CPU and GPU memory spaces as well as the data-level parallelism. In addition, the usage of zero-copy host memory, proper choice of the execution configuration, and thread organization considering memory coalescing is important to further reduce the execution time. With all these techniques, we could run QEA 3.69 times faster on average in comparison with the multi-threading CPU for the case of 0-1 knapsack problem with 30,000 items.

• Journal of Korea Game Society
• /
• v.9 no.6
• /
• pp.191-196
• /
• 2009

Volume rendering is a method which extracts meaningful information from volume data and visualizes those information. In general, since the size of volume data gets larger, it is very important to devise acceleration methods for interactive rendering speed. Min-max octree is data structure for high-speed volume rendering, however, its creation time becomes long as the data size increases. In this paper, we propose acceleration method of min-max octree generation using CUDA. Firstly, we convert one-dimensional array from volume data using space filling curve. Then we make min-max octree structures from the sequential array and apply them to acceleration of volume ray casting.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.26 no.11
• /
• pp.1012-1019
• /
• 2015

This paper shows the acceleration of iterative physical optics(IPO) for radar cross section(RCS) by using two techniques effectively. For the analysis of the multiple reflection in the cavity, IPO uses the near field method, unlike shooting and bouncing rays method which uses the geometric optics(GO). However, it is still far slower than physical optics(PO) and it is needed to accelerate the speed of IPO for practical purpose. In order to address this problem, graphic processing unit(GPU) can be applied to reduce calculation time and adaptive iterative physical optics-change rate(AIPO-CR) method is also applicable effectively to optimize iteration for acceleration of calculation.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.12
• /
• pp.1165-1176
• /
• 2010

In order to realistically visualize such participating media as cloud, smoke, and gas, the light transport process must be physically simulated inside the media. While it is known that this process is well described physically through the volume rendering equation, it usually takes a great deal of computation time for obtaining high-precision solutions. Recently, GPU-based, fast rendering methods have been proposed for the realistic simulation of participating media, however, there still remain several problems to be resolved. In this article, we describe our rendering techniques applied to enhance the performances and features of our GPU-assisted participating media renderer, and analyze how such efforts have actually improved the renderer. The presented techniques will be effectively used in volume renderers for creating various digital contents in the special effects industries.

• Journal of Korea Game Society
• /
• v.8 no.3
• /
• pp.61-68
• /
• 2008

Quadtree-based terrain visualization methods have been used in a lot of applications. However, because most procedures are performed on the CPU, the rendering speed is slow in comparison to methods using GPU. In this paper, we present a quadtree-based terrain visualization method working on the GPU with specially designed data structure, error-texture and LOD-texture, and block-based acceleration method. In preprocessing step, we calculate errors in world space and store them to error-texture. In rendering step, we examine projected errors of error-texture and choose the detail level, then store the projected errors to LOD-texture. View frustum culling is performed as block unit using the values of error-texture and LOD-texture. This method reduces CPU load and performs time consuming jobs such as LOD selection and view frustum culling.