• Proceedings of the Korean Society of Computer Information Conference
• /
• 2013.01a
• /
• pp.9-12
• /
• 2013

리눅스 기반 안드로이드 단말에서는 CFS(Completely Fair scheduler)가 사용되고 있다. 그리고 CFS는 태스크의 nice값 조절을 통해서 응용프로그램의 CPU 점유율을 제어할 수 있다. 하지만 안드로이드를 업그레이드할 때마다 수많은 태스크의 nice값을 적절하게 맞추는 일은 매우 어려운 일이다. 이러한 문제를 해결하기 위하여 안드로이드 단말은 리눅스의 cgroup(control group)을 사용하여 태스크들을 그룹으로 나눈다. 고성능과 빠른 응답 특성이 필요한 태스크들을 apps 그룹에 할당하여 높은 CPU 점유율을 보장하고, 그렇지 않은 태스크들을 background 그룹에 할당한다. 하지만 안드로이드의 버전이 업그레이드 되면서 각 그룹에 속한 태스크들에도 변화가 생긴다. 그 결과 동일하게 제작된 태스크들의 CPU 점유율이 달라지게 되고 예기치 못한 성능 저하가 발생할 수 있다. 본 연구에서는 안드로이드 버전 향상에 따라 동종 태스크들이 이전 버전에서보다 성능이 하락하는 현상의 원인을 파악하였다. 아울러 분석과 실험을 통하여 태스크의 nice 값보다 그룹 스케줄링 메커니즘이 어떻게 태스크의 CPU 점유율을 결정 짓는지 규명하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10a
• /
• pp.286-288
• /
• 2003

본 논문에서는 real-time환경에서 abnormal task를 자동 검출하여 시스템 overload 및 오 동작을 사전에 검출할 수 있는 방안을 제안한다. 본 논문에서 제안한 방안은 context switching이 발생하는 시점에서 각task들의 cpu 점유율 및 context switching 횟수를 분석하여 비정상적으로 높은 cpu 점유율을 가지는 task와 과도한 context switching을 일으켜 시스템에 overload를 주는 task들을 자동으로 검출한다. 이들 이용하여 신뢰성 있는 real-time시스템 설계 및 구현을 지원할 수 있다.

• Journal of Advanced Marine Engineering and Technology
• /
• v.34 no.6
• /
• pp.858-863
• /
• 2010

In order to enhance encoding speed in dynamic image encoding using H.264/AVC, reducing the time for motion estimation which takes a large portion of the processing time is very important. An approach using graphics processing unit(GPU) as a coprocessor to assist the central processing unit(CPU) in computing massive data, will be a way to reduce the processing time. In this paper, we present an efficient block-level parallel algorithm for the motion estimation(ME) on a computer unified device architecture(CUDA) platform developed in general-purpose computation on GPU. Experiments are carried out to verify the effectiveness of the proposed algorithm.

• Journal of Korea Game Society
• /
• v.19 no.6
• /
• pp.91-98
• /
• 2019

This study verified the memory occupancy, CPU processing speed, and average frame comparison of texture formats of ASTC and ETC in mobile Android OS. The virtual game scene was implemented as an experimental environment and built on the Android platform. Based on this, comparative verification data was extracted. ASTC has a 36% lower share of memory usage of 2D textures than ETC. CPU processing speed was 18% faster. The average frame confirmed 54 frames that was 58% higher. In the smart mobile game environment, ASTC confirmed the result of comparative advantage over ETC.

• Journal of KIISE:Computer Systems and Theory
• /
• v.36 no.5
• /
• pp.351-359
• /
• 2009

Recently with increase of data in the Internet, platform technologies that can process huge data effectively such as Google platform and Hadoop are regarded as worthy of notice. In this kind of platform, there exist network I/O overheads to send task outputs due to the MapReduce operation which is a programming model to support parallel computation in the large cluster system. In this paper, we suggest applying of TIPC (Transparent Inter-Process Communication) protocol for reducing network I/O overheads and increasing network performance in the distributed computing environments. TIPC has a lightweight protocol stack and it spends relatively less CPU time than TCP because of its simple connection establishment and logical addressing. In this paper, we analyze main features of the Hadoop-based distributed computing system, and we build an experimental model which can be used for experiments to compare the performance of various protocols. In the experimental result, TIPC has a higher bandwidth and lower CPU overheads than other protocols.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.11 no.4
• /
• pp.362-369
• /
• 2018

In this paper, we propose a procedure to multiple SNS channels automatic streaming through multiple output channels including the output channel of an external streaming server. The multiple SNS channels automatic streaming server includes an output management module for controlling the transmission of video contents to RELAY module that establish two or more output channels. In this paper, we experimented by separate with HD and FHD video using RELAY module in multiple SNS channel automatic streaming. In stream modules using RELAY module of HD video, when the publisher client and the player client and the RELAY module are 1 channel, the occupancy rate of CPU is 0.6% and the occupancy rate of heap memory is 0.3%(20 Mbyte). When the publisher client and the player client and the RELAY module are 183 channels, the occupancy rate of CPU is 99.9% and the occupancy rate of heap memory is 45.8%(3.7Gbyte). Therefore, the paper is not limited to the size of the streaming server by extending the output channel from which the video is transmitted to the output channel of the external streaming server. And a process of allocating an output channel of an external streaming server to an output channel through which an video is transmitted can be easily performed, so that an efficient output channel management can be performed even when a plurality of videos are transmitted.

• KIISE Transactions on Computing Practices
• /
• v.20 no.12
• /
• pp.623-631
• /
• 2014

Comparison of two representative deduplication file systems, LessFS and SDFS, shows that Lessfs is better in execution time and CPU utilization while SDFS is better in storage usage (around 1/8 less than general file systems). In this paper, a new system is proposed where the advantages of SDFS and Lessfs are combined. The new system uses multiple DFEs and one DSE to maintain the integrity and consistency of the data. An evaluation study to compare between Single DFE and Dual DFE indicates that the Dual DFE was better than the Single DFE. The Dual DFE reduced the CPU usage and provided fast deduplication time. This reveals that proposed system can be used to solve the problem of an increase in large data storage and power consumption.

• Journal of The Korean Association of Information Education
• /
• v.22 no.5
• /
• pp.605-611
• /
• 2018

On-Line Lectures are becoming more common due to the MOOK service and the expansion of national policy in Korea. Especially, It is being changed to new remote mixed style from traditional lecture in universities. We propose and implement a remote contents making tool with audio synchronization function based on more with less resources. To implement our proposed algorithm, we design an interactive interface to assign multiple cutting intervals and convert an input video to print a new result. In experimental, we can confirm our algorithm works properly with average performance value 9.3% cpu share ratio and 87mega byte ram usage(CPU 2.60GHz, 820*600 Area).

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.10 no.6
• /
• pp.17-22
• /
• 2010

In this paper, integral TS(Transport Stream) demultiplexer of a multi-modal receiver is proposed according to the multiple standards of European terrestrial digital broadcasting DVB-T(Digital Video Broadcasting Terrestrial), and mobile terrestrial digital broadcasting T-DMB(Terrestrial Digital Multimedia Broadcasting). This USB based integral receiver could recover the multi-modal broadcasting audios by memory sharing technique which was utilized to decrease the load by the control of streaming multi-modal broadcasting. As a result of performance analysis for a proposed integral TS demultiplexer, the CPU occupational efficiency of windows based integral demulitiplexing is improved compared with DVB-T, and T-DMB respectively.

• The KIPS Transactions:PartC
• /
• v.18C no.3
• /
• pp.157-166
• /
• 2011

The virtualization layer is executed in higher authority layer than kernel layer and suitable for monitoring operating systems. However, existing virtualization monitoring systems provide simple information about the usage rate of CPU or memory. In this paper, the monitoring system using full virtualization technique is proposed, which can monitor virtual machine's dynamic kernel object as memory, register, GDT, IDT and system call table. To verify the monitoring system, the proposed system was implemented based on KVM(Kernel-based Virtual Machine) with full virtualization that is directly applied to linux kernel without any modification. The proposed system consists of KvmAccess module to access KVM's internal object and API to provide other external modules with monitoring result. In experiments, the CPU utilization for monitoring operations in the proposed monitering system is 0.35% when the system is monitored with 1-second period. The proposed monitoring system has a little performance degradation.