• 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.

• ETRI Journal
• /
• v.45 no.6
• /
• 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.

• The Journal of the Korea Contents Association
• /
• v.22 no.6
• /
• pp.56-68
• /
• 2022

Recently, as the demand for real-time processing increases, studies on a dynamic graph that changes over time has been actively done. There is a connected components processing algorithm as one of the algorithms for analyzing dynamic graphs. GPUs are suitable for large-scale graph calculations due to their high memory bandwidth and computational performance. However, when computing the connected components of a dynamic graph using the GPU, frequent data exchange occurs between the CPU and the GPU during real graph processing due to the limited memory of the GPU. The proposed scheme utilizes the Weighted-Quick-Union algorithm to process large-scale graphs on the GPU. It supports fast connected components computation by applying the size to the connected component label. It computes the connected component by determining the parts to be recalculated and minimizing the data to be transmitted to the GPU. In addition, we propose a processing structure in which the GPU and the CPU execute asynchronously to reduce the data transfer time between GPU and CPU. We show the excellence of the proposed scheme through performance evaluation using real dataset.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.4
• /
• pp.108-116
• /
• 2013

Many algorithms have been introduced to improve performance by using massively parallel systems, which consist of several hundreds of processors. A typical example is a GPU system of many processors which uses shared memory. In the case of image filtering algorithms, which make references to neighboring points, the shared memory helps improve performance by frequently accessing adjacent pixels. However, using shared memory requires rewriting the existing codes and consequently results in complexity of the codes. Recent GPU systems support both L1 and L2 cache along with shared memory. Since the L1 cache memory is located in the same area as the shared memory, the improvement of performance is predictable by using the cache memory. In this paper, the performance of cache and shared memory were compared. In conclusion, the performance of cache-based algorithm is very similar to the one of shared memory. The complexity of the code appearing in a shared memory system, however, is resolved with the cache-based algorithm.

• Spatial Information Research
• /
• v.20 no.6
• /
• pp.139-144
• /
• 2012

The R*-tree manages hierarchical nodes for efficient access of spatial data. We propose a method that maintains partial nodes of R*-tree in the GPU memory to improve efficiency using parallel processing. The proposed method attempts to load as many nodes as possible to the GPU memory. The new nodes are inserted to manage the rest of R*-tree nodes in the main memory. The experimental result shows that the proposed method is more efficient than the main memory based R*-tree.

• Journal of Digital Contents Society
• /
• v.9 no.3
• /
• pp.371-377
• /
• 2008

Recently the performance of GPU is increasing much faster compared to GPU and GPU is used for various application programs. In this paper, MPEG-2 Decoder is implemented based on a GPU programming language, CG. The proposed methodology is to perform block rendering with texture data according to video standard with very high parallelism by using the pipeline of GPU which is a stream processing structure. To reduce the data bandwidth between system memory and GPU, local memory is used for graphic card. According to the experiment, the proposed scheme shows performance improvement by more than 2 times compared to CPU based scheme.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.8
• /
• pp.778-785
• /
• 2009

This paper presents a Gabor texture feature extraction method for classification of discolored Metal pad images using GPU(Graphics Processing Unit). The proposed algorithm extracts the texture information using Gabor filters and constructs a pattern map using the extracted information. Finally, the golden pad images are classified by utilizing the feature vectors which are extracted from the constructed pattern map. In order to evaluate the performance of the Gabor texture feature extraction algorithm based on GPU, a sequential processing and parallel processing using OpenMP in CPU of this algorithm were adopted. Also, the proposed algorithm was implemented by using Global memory and Shared memory in GPU. The experimental results were demonstrated that the method using Shared memory in GPU provides the best performance. For evaluating the effectiveness of extracted Gabor texture features, an experimental validation has been conducted on a database of 20 Metal pad images and the experiment has shown no mis-classification.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.9
• /
• pp.1-12
• /
• 2021

In this paper, we propose a tool called ARCAV (Atomatic Recovery of CUDA Atomicity violation) to automatically repair atomicity violations in GPU (Graphics Processing Unit) program. ARCAV monitors information of every barrier and memory to make actual memory writes occur at the end of the barrier region or to make the program execute barrier region again. Existing methods do not repair atomicity violations but only detect the atomicity violations in GPU programs because GPU programs generally do not support lock and sleep instructions which are necessary for repairing the atomicity violations. Proposed ARCAV is designed for GPU execution model. ARCAV detects and repairs four patterns of atomicity violations which represent real-world cases. Moreover, ARCAV is independent of memory hierarchy and thread configuration. Our experiments show that the performance of ARCAV is stable regardless of the number of threads or blocks. The overhead of ARCAV is evaluated using four real-world kernels, and its slowdown is 2.1x, in average, of native execution time.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.11a
• /
• pp.197-199
• /
• 2012

Recently, a lot of general-purpose application programs in addition to graphic applications have been parallelized for boosting their performance using Graphic Processing Unit (GPU)'s excellent floating-point performance. In order to maximize the application performance on GPUs, optimizing the memory hierarchy and the on-chip caches such as the shared memory is essential. In this paper, we propose techniques to optimize the shared memory, and verify its effectiveness using a pattern matching application program.

• Journal of Computing Science and Engineering
• /
• v.11 no.2
• /
• pp.69-77
• /
• 2017

Recent GPUs have adopted cache memory to benefit general-purpose GPU (GPGPU) programs. However, unlike CPU programs, GPGPU programs typically have considerably less temporal/spatial locality. Moreover, the L1 data cache is used by many threads that access a data size typically considerably larger than the L1 cache, making it critical to bypass L1 data cache intelligently to enhance GPU cache performance. In this paper, we examine GPU cache access behavior and propose a simple hardware-based GPU cache bypassing method that can be applied to GPU applications without recompiling programs. Moreover, we introduce a hybrid method that integrates static profiling information and hardware-based bypassing to further enhance performance. Our experimental results reveal that hardware-based cache bypassing can boost performance for most benchmarks, and the hybrid method can achieve performance comparable to state-of-the-art compiler-based bypassing with considerably less profiling cost.