• 한국컴퓨터정보학회논문지
• /
• 제27권2호
• /
• pp.1-8
• /
• 2022

본 논문에서는 개미 집단 시스템(ant colony system)을 통한 순회 외판원 문제(traveling salesman problem)를 효과적으로 해결하기 위해 GPU 기반 병렬 알고리즘을 설계 구현하였다. TSP에서 동시에 수백 또는 수천의 탐색 여정(tour)을 생성하는 반복 과정을 GPU의 작업 병렬성을 활용하여 처리성능을 개선하고, 페로몬 자취 데이터의 업데이트 과정은 32x32의 쓰레드 블럭을 사용하여 데이터 병렬성을 적극 활용하였다. 특히 다중 쓰레드의 메모리 동시 접근을 통해 연속 메모리공간의 병합 접근 효과와 공유 메모리의 동시 접근을 지원하였다. 본 실험은 TSPLIB에서 제공되는 127개부터 1002개에 이르는 도시 데이터를 사용하였고, Intel Core i9-9900K CPU와 Nvidia Titan RTX 시스템을 사용하여 순차 알고리즘과 병렬 알고리즘의 성능을 비교하였다. GPU 병렬화에 의한 성능 향상은 약 10.13~11.37배의 성능 개선 효과를 보였다.

• Journal of Positioning, Navigation, and Timing
• /
• 제9권4호
• /
• pp.337-345
• /
• 2020

In this paper, an efficient signal tracking method to simultaneously track both GPS L1 C/A and Galileo E1B CBOC(6,1,1/11) using a low cost GPU is proposed. In the existing method that each GNSS signal is processed within 1 ms, more than 2 ms processing time is required in GPU to process 4 ms CBOC signal. It means that real time operation is possible if only Galileo E1B CBOC signal is concerned. But when both GPS C/A and Galileo CBOC is required, it cannot process GPS C/A signal in real time. To process 1 ms GPS C/A and 4 ms Galileo CBOC signal in real time, 4 ms Galileo CBOC signal is divided into 4 by 1 ms signal block in the proposed method. Specially, a buffer that simultaneously manages 1 ms and 4 ms signals is designed. In addition, a module that accumulates the 1 ms correlation value of the Galileo CBOC by 4 ms and passes it to the PLL and DLL is implemented. The operation and performance are evaluated with real measurements in the GPU based SDR. The experimental results show that tracking of more than 16 satellites of GPS C/A and Galileo E1B is possible using the proposed method.

• Spatial Information Research
• /
• 제23권2호
• /
• pp.1-9
• /
• 2015

그래픽 처리 장치(GPU)는 내부에 대량의 산술 논리 연산 장치(ALU)를 보유하고 있다. 대량의 ALU는 병렬 처리를 위해 이용될 수 있으므로, GPU는 효율적인 데이터 처리를 제공한다. 공간 데이터를 지도상에 표현하기 위하여 지리학적 좌표가 필요하다. 좌표들은 측지경도와 측지위도의 형태로 저장된다. 데카르트 좌표계로 구성된 지도를 표현하기 위하여 측지경도와 측지위도는 국제 횡단 메르카토르 좌표계(UTM)로 전환돼야 한다. 좌표계 변환 과정과 변환된 좌표를 화면상에 표현하기 위한 렌더링 과정은 복잡한 부동 소수점 계산이 필요하다. 본 논문에서는 성능 향상을 위해 GPU를 활용한 좌표변환 과정과 렌더링 과정을 병렬적으로 처리하는 기법을 제안한다. 대용량 공간 데이터는 파일로 디스크 내에 저장된다. 대용량 공간 데이터를 효율적으로 처리하기 위하여 공간 데이터 파일들을 하나의 대용량 파일로 병합하고 Memory Mapped File 기법을 활용하여 파일에 접근하는 기법을 제안한다. 본 논문에서는 TIGER/Line 데이터를 활용하여 747,302,971개의 점으로 구성된 공간 데이터의 좌표 변환 및 렌더링 처리 과정을 GPU를 활용하여 병렬로 수행하는 연구를 진행한다. CPU를 이용하여 좌표변환 과정 결과와 렌더링 처리 과정 결과를 비교하여 속도 향상 정도에 대한 결과를 제시한다.

• Journal of information and communication convergence engineering
• /
• 제19권4호
• /
• pp.257-262
• /
• 2021

Providing web services to users has become expensive in recent times. For better web services, a web server is provided with high-performance technology. To achieve great web service experiences, tools such as general-purpose graphics processing units (GPGPUs), artificial intelligence, high-performance computing, and three-dimensional simulation are widely used. However, graphics processing units (GPUs) are used in high-speed operations and have limited general applications. In this study, we developed a task queue in a GPU to improve the performance of a web service using a multiprocessor and studied how to receive and process user requests in bulk. We propose the use of a GPGPU-based task queue to process user requests more than GPGPU based a central processing unit thread, and to process more GPU threads on task queue at about 136% to 233%, and proved that the proposed method is effective for web service.

• 한국정보통신학회논문지
• /
• 제17권5호
• /
• pp.1119-1128
• /
• 2013

본 논문에서는 얼굴을 검출하고 GPU 기반으로 얼굴을 고속으로 추적하는 알고리즘을 제안하였다. 얼굴 검출에서는 깊이영상과 RGB영상을 사용하고, 기존의 방법인 Adaboost을 이용하지만 움직임 영역과 피부색 영역을 이용하여 Adaboost의 입력영상을 제한하여 얼굴을 검출하였다. 얼굴 검출과는 다르게 얼굴 추적은 깊이 정보만을 사용하였다. 기본적으로 얼굴 추적에서는 템플릿과 매칭 된 블록을 찾는 템플릿 매칭 방법을 사용하였다. 또한 고속으로 얼굴을 추적하기 위해서 GPU를 이용하여 템플릿 매칭을 병렬하여 연산하였다. 실험결과 CPU와 GPU을 비교 하였을 때 GPU 수행속도가 최대 49배까지 향상되는 것을 확인하였다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2005년도 추계종합학술대회
• /
• pp.1003-1006
• /
• 2005

Vertex shader of GPU in personal computer is advanced in functions as to be half of traditional fixed T&L functions. And, capacity of memory for saving resources to process instructions is unlimited. GPU that can be programmed by programmer is needed for mobile system as well as personal computer. In this paper, we implement software virtual machine for vertex shader using C++ Language. Our goal is designing hardware GPU that can apply to mobile system. The virtual machine consists of nVidia GPU instructions. Input Data to virtual machine is generated by Microsoft fxc compiler. That is to say, Input Data is compiled shader program written in HLSL, Cg, or ASM. The virtual machine will be a reference model for designing hardware GPU and can be used for Testbed to test added or modified instruction.

• 대한임베디드공학회논문지
• /
• 제11권2호
• /
• pp.57-65
• /
• 2016

In this paper, we propose a system for efficient use of shared memory between CPU and GPU. The system, called Fusion Architecture, assures consistency of the shared memory and minimizes cache misses that frequently occurs on Heterogeneous System Architecture or Unified Virtual Memory based systems. It also maximizes the performance for memory intensive jobs by efficient allocation of GPU cores. To test between architectures on various scenarios, we introduce the Fusion Architecture Analyzer, which compares OpenMP, OpenCL, CUDA, and the proposed architecture in terms of memory overhead and process time. As a result, Proposed fusion architectures show that the Fusion Architecture runs benchmarks 55% faster and reduces memory overheads by 220% in average.

• Journal of Information Processing Systems
• /
• 제7권1호
• /
• pp.63-74
• /
• 2011

As a powerful and flexible processor, the Graphic Processing Unit (GPU) can offer a great faculty in solving many high-performance computing applications. Sweep3D, which simulates a single group time-independent discrete ordinates (Sn) neutron transport deterministically on 3D Cartesian geometry space, represents the key part of a real ASCI application. The wavefront process for parallel computation in Sweep3D limits the concurrent threads on the GPU. In this paper, we present multi-dimensional optimization methods for Sweep3D, which can be efficiently implemented on the finegrained parallel architecture of the GPU. Our results show that the overall performance of Sweep3D on the CPU-GPU hybrid platform can be improved up to 4.38 times as compared to the CPU-based implementation.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2014년도 제49차 동계학술대회논문집 22권1호
• /
• pp.5-8
• /
• 2014

This paper presents a parallel approach of formant synthesis method for haegeum on graphics processing units (GPU) using spectral modeling. Spectral modeling synthesis (SMS) is a technique that models time-varying spectra as a combination of sinusoids and a time-varying filtered noise component. A second-order digital resonator by the impulse-invariant transform (IIT) is applied to generate deterministic components and the results are band-pass filtered to adjust magnitude. The noise is calculated by first generating the sinusoids with formant synthesis, subtracting them from the original sound, and then removing some harmonics remained. The synthesized sounds are consequently by adding sinusoids, which are shown to be similar to the original Haegeum sounds. Furthermore, GPU accelerates the synthesis process enabling- real time music synthesis system development, supporting more sound effect, and multiple musical sound compositions.

• 예술인문사회 융합 멀티미디어 논문지
• /
• 제4권2호
• /
• pp.369-378
• /
• 2014

The Reed-Muller code is one of the efficient algorithms for multiple bit error correction, however, its high-computation requirement inherent in the decoding process prohibits its use in practical applications. To solve this problem, this paper proposes a graphics processing unit (GPU)-based parallel error control approach using Reed-Muller R(r, m) coding for real-time wireless communication systems. GPU offers a high-throughput parallel computing platform that can achieve the desired high-performance decoding by exploiting massive parallelism inherent in the algorithm. In addition, we compare the performance of the GPU-based approach with the equivalent sequential approach that runs on the traditional CPU. The experimental results indicate that the proposed GPU-based approach exceedingly outperforms the sequential approach in terms of execution time, yielding over 70× speedup.