• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.4
• /
• pp.77-82
• /
• 2024

Recent advancements in artificial intelligence (AI) technology have led to an increase in the implementation of AI applications in mobile environments. However, due to the limited resources in mobile devices compared to desktops and servers, there is growing interest in research aimed at efficiently executing AI workloads on mobile platforms. While most studies focus on offloading to edge or cloud solutions to mitigate computing resource constraints, research on the characteristics of file I/O related to storage access in mobile settings remains underexplored. This paper analyzes file I/O traces generated during the execution of deep learning applications in mobile environments and investigates how they differ from traditional mobile workloads. We anticipate that the findings of this study will be utilized to design future smartphone system software more efficiently, considering the file access characteristics of deep learning.

• The KIPS Transactions:PartD
• /
• v.11D no.5
• /
• pp.1041-1048
• /
• 2004

Database logs contain various information on database operations, but they are used to recover database systems from failures generally. This paper proposes a log analysis tool that provides useful information for database tuning. This tool provides users with information on work-load organization, database schemas, and resources usages of queries. This paper describes the tool in views of its architecture, functions, implementation, and verification. The tool is verified by running the TPC-W benchmark, and representative analysis results are also presented.

• KIISE Transactions on Computing Practices
• /
• v.23 no.3
• /
• pp.165-170
• /
• 2017

Many studies have utilized GPGPU (General-Purpose Graphic Processing Unit) and its high computing power to compute complex tasks. The characteristics of GPGPU programs necessitate the operations of memory copy between the host and device. A high latency period can affect the performance of the program. Thus, it is required to significantly improve the performance of GPGPU programs by optimizations. By executing multiple GPGPU programs simultaneously, the latency hiding effect of memory copy is achieved by overlapping the memory copy and computing operations in GPGPU. This paper presents the results of analyzing the latency hiding effect for memory copy operations. Furthermore, we propose a performance anticipation model and an algorithm for the limitations of using pinned memory, and show that the use of the proposed algorithm results in a 41% performance increase.

• Journal of KIISE
• /
• v.41 no.10
• /
• pp.832-837
• /
• 2014

Application loading speed can be improved by timely prefetching disk blocks likely to be needed by an application. However, existing prefetchers -- if they are not specialized to a particular application -- incur high overheads and are poor at identifying the blocks that will actually be required. There are many sequences in which blocks may be needed and, even if two access sequences are identical, block tracing and access timings can be affected significantly by the state of the buffer cache. We propose a new application-independent software-based prefetching technique, in which breakpoints are inserted at appropriate places in an application to collect the information on correlations between the blocks and to prefetch the potential blocks ahead of their schedule based on it. Experiments on an HDD-based desktop PC demonstrated an average 30% reduction in application launch time and 15% in general I/O, while reducing the wasted overhead.

• Annual Conference of KIPS
• /
• 2022.11a
• /
• pp.147-149
• /
• 2022

PIM은 CPU와 메모리 간의 데이터 버스 오버헤드를 완화하기 위해서 메모리 내부에 프로세서를 가지며 낮은 데이터 재사용성을 가지는 데이터 집약형 워크로드에서 지연과 에너지 관점에서 장점을 가진다. 본 논문은 UPMEM사의 PIM을 이용하여 HPC분야에서 자주 사용되는 행렬 연산인 GEMV, SpMV의 벤치마크 구현을 분석하고 성능 분석을 통해 CPU 대비 가지는 장단점에 대해서 논하였다.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.2
• /
• pp.123-130
• /
• 2013

As size of genomic data is increasing rapidly, the needs for high-performance computing system to process and store genomic data is also increasing. In this paper, we captured I/O trace of a system which analyzed 500 million sequence reads data in Genome analysis pipeline for 86 hours. The workload created 630 file with size of 1031.7 Gbyte and deleted 535 file with size of 91.4 GByte. What is interesting in this workload is that 80% of all accesses are from only two files among 654 files in the system. Size of read and write request in the workload was larger than 512 KByte and 1 Mbyte, respectively. Majority of read write operations show random and sequential patterns, respectively. Throughput and bandwidth observed in each processing phase was different from each other.

• The Journal of the Korea Contents Association
• /
• v.21 no.9
• /
• pp.198-208
• /
• 2021

ArangoDB is a NoSQL database system that has been popularly utilized in many applications for storing large amounts of data. In order to apply a new NoSQL database system such as ArangoDB, to real work environments we need a benchmarking system that can evaluate its performance. In this paper, we design and implement a ArangoDB based benchmarking system that measures a kernel level performance well as an application level performance. We partially modify YCSB to measure the performance of a NoSQL database system in the cluster environment. We also define three real-world workload types by analyzing the existing materials. We prove the feasibility of the proposed system through the benchmarking of three workload types. We derive available workloads in ArangoDB and show that performance at the kernel layer as well as the application layer can be visualized through benchmarking of three workload types. It is expected that applicability and risk reviews will be possible through benchmarking of this system in environments that need to transfer data from the existing database engine to ArangoDB.

• Annual Conference of KIPS
• /
• 2021.11a
• /
• pp.32-35
• /
• 2021

단일 가상화 서비스 플랫폼은 메모리 및 컴퓨팅 집약적 워크로드를 수행하기 위한 고성능 시스템 환경의 신속한 구축을 지원하는 클라우드 기반의 소프트웨어 정의 서버를 위한 핵심 기술이다. 본 연구는 다수의 물리 노드를 통합하여 하나의 고성능 단일가상서버로 구성하기 위해 개발된 HCC 단일 가상화 서비스 플랫폼에서 대용량 데이터 처리 및 대규모 연산이 필요한 NGS 기반 농생명유전체 조립 프로그램과 이상 기상의 탐지 분석을 위한 GOES 위성자료 전처리 프로그램을 시험하여 활용 적합성을 검증하였다.

• Convergence Security Journal
• /
• v.13 no.4
• /
• pp.93-100
• /
• 2013

In 2012, According to 'Cisco Global Cloud Index 2011-2016', the Cisco company forecasted that global data center traffic will nearly quadruple and cloud traffic will nearly sextuple by 2016. Such a rapid growing of traffic is caused by traffic inside the data center and cloud computing workloads. In 2010, the Ministry of National Defense decided to build a Mega Center including the cloud computing technology by 2014, as part of the '2012 Information Service Plan', which is now underway. One of the factors to consider is cloud computing traffic to build a Mega Center. Since the K-defense cloud computing system is built, K-defense cloud computing traffic will increase steadily. This paper, analyzed the availability of K-defense cloud computing service with the K-defense cloud computing traffic increasing using K-Defense cloud computing test system and CloudAnalyst simulation tool. Created 3 scenarios and Simulated with these scenarios, the results are derived that the availability of K-defense cloud computing test system is fulfilled, even cloud workloads are increased as muh as forecasted cloud traffic growth from now until 2016.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.8
• /
• pp.1-10
• /
• 2023

Recently, NAND flash memories have replaced magnetic disks due to non-volatility, high capacity and high resistance, in various computer systems but it has disadvantages which are the limited lifespan and imbalanced operation latency. Therefore, many page replacement policies have been studied to overcome the disadvantages of NAND flash memories. Although it is clear that these policies reflect execution characteristics of various environments and applications, researches on the foundation-policy decision for disk buffer management are insufficient. Thus, in this paper, we propose a foundation-policy recommendation model, called FRM for effectively utilizing NAND flash memories. FRM proposes a suitable page replacement policy by classifying and analyzing characteristics of workloads through machine learning. As an implementation case, we introduce FRM with a disk buffer management policy and in experiment results, prediction accuracy and weighted average of FRM shows 92.85% and 88.97%, by training dataset and validation dataset for foundation disk buffer management policy, respectively.