• ETRI Journal
• /
• 제45권6호
• /
• pp.1035-1045
• /
• 2023

We propose a novel graphics processing unit (GPU) algorithm that can handle a large-scale 3D fast Fourier transform (i.e., 3D-FFT) problem whose data size is larger than the GPU's memory. A 1D FFT-based 3D-FFT computational approach is used to solve the limited device memory issue. Moreover, to reduce the communication overhead between the CPU and GPU, we propose a 3D data-transposition method that converts the target 1D vector into a contiguous memory layout and improves data transfer efficiency. The transposed data are communicated between the host and device memories efficiently through the pinned buffer and multiple streams. We apply our method to various large-scale benchmarks and compare its performance with the state-of-the-art multicore CPU FFT library (i.e., fastest Fourier transform in the West [FFTW]) and a prior GPU-based 3D-FFT algorithm. Our method achieves a higher performance (up to 2.89 times) than FFTW; it yields more performance gaps as the data size increases. The performance of the prior GPU algorithm decreases considerably in massive-scale problems, whereas our method's performance is stable.

• 한국항행학회논문지
• /
• 제22권5호
• /
• pp.429-435
• /
• 2018

레이다 시스템의 경우, 타겟의 거리와 속도를 추출하기 위해 FFT (fast Fourier transform) 연산이 필수적으로 요구되며, 실시간 구현을 위해 고속으로 동작하는 FFT 프로세서의 설계가 필요하다. 고속 FFT 프로세서를 위한 하드웨어 구조로 완전 셔플 네트워크 (perfect shuffle network) 구조가 적합하며, 특히 초고속 연산을 위해 radix-4 기반의 이중 완전 셔플 네트워크 (twice perfect shuffle network) 구조가 가장 적절하고 볼 수 있다. 더불어, 다양한 속도 해상도를 요구하는 레이다 응용을 고려할 때, FFT 프로세서는 가변길이 FFT 연산을 지원할 필요가 있다. 이에 본 논문에서는 8~1024 포인트의 가변 길이 연산을 지원하는 이중 완전 셔플 네트워크 기반의 FFT 알고리즘을 제안하였으며, 이의 하드웨어 구조 설계 및 구현 결과를 제시한다. 제안된 FFT 프로세서는 HDL (hardware description language)을 활용하여 RTL (register transfer level) 설계가 수행되었으며, $0.65{\mu}m$ CMOS 공정을 활용하여 논리 합성한 결과, 총 3,293K개의 논리 게이트로 구현 가능함을 확인 할 수 있었다.

• 한국연소학회지
• /
• 제22권2호
• /
• pp.1-8
• /
• 2017

To evaluate the combustion instability of a gas turbine combustor, the DMD technique was applied. The mode frequency results for each fuel composition were compared with FFT(Fast Fourier Transform) results. The damping coefficient, which is a quantitative parameter for combustion instability, was evaluated for 5 experimental cases. The flame transfer function (FTF) was calculated in the most unstable test case. In deriving the FTF, gain and phase were calculated using DMD technique. As a result of the analysis of the OH radical perturbation of the DMD, the heat release fluctuation was the highest at 100 Hz, at which the highest value of gain is observed. The frequency of FFT and FTF were different. In order to clarify the reason for this, FTF for various resonance frequencies was performed and it shows that the pattern of gain was similar to FFT.

• 한국항행학회논문지
• /
• 제24권2호
• /
• pp.85-91
• /
• 2020

최근 다양한 환경에서 무인기를 효율적으로 운용하기 위한 목적으로 멀티모드 레이다 시스템이 고안되었으며, 이는 PD (pulse Doppler) 방식과 FMCW (frequency modulated continuous wave) 방식을 통합하여 활용할 수 있다는 장점을 가진다. 멀티모드 레이다 시스템의 하드웨어 구조의 경우 FFT (fast Fourier transform) 프로세서와 IFFT (inverse fast Fourier transform) 프로세서가 필수적이지만, FFT 프로세서는 큰 복잡도를 갖는 구조 중 하나로 FFT 프로세서의 복잡도를 감소시키는 방향으로의 구조 설계가 필요하다. 또한, 다양한 거리 해상도를 요구하는 레이다 응용 환경을 고려했을 때, FFT 프로세서는 가변 길이의 연산을 지원할 필요가 있다. 이에 본 논문에서는 멀티모드 레이다 신호처리 프로세서 거리 추정부의 FFT 프로세서와 IFFT 프로세서를 16~1024 포인트의 가변 길이 연산을 지원하는 단일 FFT 프로세서의 하드웨어로 설계하여 제안한다. 제안된 FFT 프로세서는 MATLAB 기반 알고리즘 설계를 수행한 뒤, 그 결과를 토대로 Verilog-HDL (hardware description language)을 활용하여 RTL (register transfer level) 설계가 수행되었으며, 논리 합성 결과 총 총 7,452개의 logic elements, 5,116개의 registers로 구현 가능함을 확인하였다.

• 천문학회지
• /
• 제48권1호
• /
• pp.67-73
• /
• 2015

We present a novel method to implement time-delayed propagation of radiation fields in cosmological radiative transfer simulations. Time-delayed propagation of radiation fields requires construction of retarded-time fields by tracking the location and lifetime of radiation sources along the corresponding light-cones. Cosmological radiative transfer simulations have, until now, ignored this "light-cone effect" or implemented ray-tracing methods that are computationally demanding. We show that radiative transfer calculation of the time-delayed fields can be easily achieved in numerical simulations when periodic boundary conditions are used, by calculating the time-discretized retarded-time Green's function using the Fast Fourier Transform (FFT) method and convolving it with the source distribution. We also present a direct application of this method to the long-range radiation field of Lyman-Werner band photons, which is important in the high-redshift astrophysics with first stars.

• 연구논문집
• /
• 통권31호
• /
• pp.77-88
• /
• 2001

In this paper, the critical vibration limits of a spindle unit for a high speed ball pen tip processing machine are studied. The vibration of the spindle is measured by FFT, and the influence of the vibration amplitude due to unbalance, bearing deflect, bite, and timing belts tension is analyzed. The critical vibration limits of the spindle is determined by the X, and Z directional vibrations of the spindle. Both FET and RMS values can be used to analyze the vibration characteristics of the spindle. From experimental results the limit line can be drawn for the spindle. The RMS value for the vibration limit is 3 G.

• 한국도로학회논문집
• /
• 제15권4호
• /
• pp.31-42
• /
• 2013

PURPOSES : The purpose of this study is to eliminate the noise of the vehicle after measuring the friction noise obtained from the NCPX (Noble Close ProXimity) method. The pure friction noise between the tire and road pavement could be determined from filtering the compositeness of sound and the influence of the vehicle noise. METHODS: The noise magnitude could be determined by analyzing the sound pressure level (SPL) and sound power level (PWL) along with the noise frequency of a FFT (Fast Fourier Transform) analysis as well as CPB (Constant Percentage Bandwidth) analysis. RESULTS: When the test for measuring the friction noise originated somewhere between tire and road pavement is performed with NCPX method, it must be fulfilled by attaching the surface microphone near the tire. In this condition, the surface microphone can measure the friction noise occurred at between tire and pavement, the chassis noise from the engine and power transfer units, the fluctuating aerodynamic noise, and the turbulence noise directly affected to the surface microphone. By using the NCPX method, the noise occurred at the vehicle must be eliminated for measuring the friction noise between tire and pavement from the traffic noise. CONCLUSIONS: The vehicle's testing engine noise depends on the vehicle and road types. The effect of vehicle's engine noise is less than the friction noise occurred at between tire and pavement at less than 1% effect.

• 한국광학회지
• /
• 제33권3호
• /
• pp.106-112
• /
• 2022

이중 링 add/drop 필터와 all-pass 지연 필터로 구성된 이차원 optical code division multiple access (OCDMA) 인코더/디코더를 제안하고, 설계 예시 및 수치해석을 통하여 실현 가능성을 확인하였다. 제안된 OCDMA 인코더/디코더의 칩 면적은 지연 도파로를 사용하는 기존 OCDMA 인코더/디코더에 비하여 1/3 정도로 줄어든다. 제안된 소자의 성능을 모델링하기 위하여 고속 푸리에 변환(fast Fourier transform, FFT) 및 전달 행렬 기법을 사용하였다. 정확한 코드로 디코딩된 펄스의 중심에서 자기상관 피크 값은 어긋난 파장 호핑 코드 및 스펙트럼 위상 코드로 디코딩된 경우의 최대 교차상관 레벨에 비하여 3배 이상으로 관측되었다. 이를 통하여 forward error correction (FEC) 한계에 해당하는 10^-3 이하의 비트 에러 오율을 얻을 수 있음을 알 수 있다.

• Journal of Multimedia Information System
• /
• 제4권4호
• /
• pp.225-232
• /
• 2017

Understanding actions in videos is an important task. It helps in finding the anomalies present in videos such as fights. Detection of fights becomes more crucial when it comes to sports. This paper focuses on finding fight scenes in Hockey sport videos using blur & radon transform and convolutional neural networks (CNNs). First, the local motion within the video frames has been extracted using blur information. Next, fast fourier and radon transform have been applied on the local motion. The video frames with fight scene have been identified using transfer learning with the help of pre-trained deep learning model VGG-Net. Finally, a comparison of the methodology has been performed using feed forward neural networks. Accuracies of 56.00% and 75.00% have been achieved using feed forward neural network and VGG16-Net, respectively.

• Fibers and Polymers
• /
• 제5권3호
• /
• pp.177-181
• /
• 2004

Current efforts to establish links between geometrical features and mechanical performance of nonwoven fabrics in general, and of point-bonded (spot-bonded) nonwovens in particular has been made using the measurements of Fiber Orientation Distribution Function (ODF) and tensile modulus which occurs during controlled-deformation experiments. Image analysis technique (using the Fast Fourier Transform) was used to quantify the fiber orientation distribution. The results suggest that, within a typical window of processing conditions, the fiber orientation has a significant influence on the anisotropical behavior of nonwoven. The data also suggest that mechanical anisotropy of thermally point-bonded nonwovens is likely to be governed by different load transfer mechanism according to the applied macroscopic tensile direction.