• The KIPS Transactions:PartA
• /
• v.12A no.1 s.91
• /
• pp.41-52
• /
• 2005

Memory reference instruction caused by cache miss is the critical factor that limits the processing power of processor. Cache prefetching technique is an effective way to reduce the latency due to memory access. However, excessively aggressive prefetch leads to cache pollution and finally to cancel out the advantage of prefetch. In this study, an active prefetch filtering scheme is introduced which dynamically decides whether to commence prefetching after referring a filtering table to reduce the cache pollution due to unnecessary prefetches. For the precision filtering, an evicted address referencing scheme has been proposed where the filter directly compares the current prefetch address with previous unnecessary prefetch addresses stored in filtering table. Moreover, a small sized exclusive prefetch cache has been introduced to increase the amount of eviction of unnecessarily prefetched addresses to enhance the accuracy of dynamic filtering. The exclusive prefetch cache also prevents useful demand data from being pushed out by prefetched data, while the evicted address direct referencing scheme enables the prefetch cache to keep most of useful prefetch data within its small size. Experimental results from commonly used general and multimedia benchmarks show that the average cache miss ratio has been decreased by $13.3{\%}$ by virtue of enhanced filtering accuracy compared with conventional schemes.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2003.11a
• /
• pp.417-420
• /
• 2003

요즘은 여러 멀티미디어 데이터가 기존의 PC에서 재생되는 것이 아니라 핸드폰, PDA와 같은 이동단말기에서 재생되기가 요구되고 있다 멀티미디어 데이터의 용량이 크고 시간에 종속적인 연속 미디어이기 때문에 데이터의 일부론 미리 가져와서 재생하는 Prefetch기법이 필요하다. 영역 분할 기법과 재생가능성 분석기법을 통해서 멀티미디어 프리젠테이션의 재생가능 여부를 분석하고 재생이 불가능한 프리젠테이션은 EPOB지정을 이용한 Prefetch기법을 적용한 후 재생가능 여부를 판단한다. Prefetch하는 시점과 데이터의 양을 지정함으로서 원활한 멀티미디어 프리젠테이션을 재생 할 수 있다.

• Journal of KIISE:Software and Applications
• /
• v.26 no.12
• /
• pp.1404-1417
• /
• 1999

웹과 같은 분산 환경에서는, 웹 브라우저 상에서 SQL 형식의 공간 질의를 수행시키는 것과 또한 서버로부터 그 질의 결과를 보는 것이 가능하다. 그러나, 격자 이미지(raster image)와 같은 대용량 공간 데이타를 포함하는 질의 결과를 웹 브라우징할 때 발생하는 많은 문제점들 중에서, 사용자 응답 시간의 지연은 매우 중요한 문제이다. 본 논문에서는 사용자의 재요청(callback) 접근 패턴이 공간적 근접성(spatial locality)을 따른다는 가정하에서의, 사용자 응답 시간을 최소화하기 위한 새로운 프리페치(prefetch) 전략에 대해서 서술한다. 본 논문의 프리페치 전략은 다음과 같이 요약될 수 있다. 첫째, 프리페치 알고리즘은 사용자의 접근 패턴을 잘 반영하는 힐버트 곡선(Hilbert-curve) 모델을 바탕으로 한다. 둘째, 프리페치 전송 비용을 줄이기 위해서 사용자의 재요청 시간 간격(think time)을 이용한다. 본 논문에서는, 힐버트 곡선을 이용한 프리페치 전략의 성능 평가를 위해서 다양한 실험을 하였으며, 그 결과로 프리페치를 하지 않는 방식보다 높은 성능 향상이 있음을 보인다.Abstract In distributed environment(e.g., WWW), it would be possible for the users to submit SQL-like spatial queries and to see their query results from the server on the Web browser. However, of many obstacles which result from browsing query results including large spatial data such as raster image, the delay of user response time is very critical. In this paper we present a new prefetch policy which can alleviate user response time on the assumption that user's callback access pattern has spatial locality. Our prefetch policy can be summerized as follows: 1) our prefetch algorithm is based on the Hibert-curve model which well replects user's access pattern, and 2) it utilizes user's callback interval to reduce prefetch network transmission cost. In this paper we conducted diverse experiments to show that our prefetch policy achieves higher performance improvement compared to other non-prefetch methods.

• Electronics and Telecommunications Trends
• /
• v.13 no.4 s.52
• /
• pp.59-69
• /
• 1998

The terminology of data prefetching is introduced, which includes stride, repeat distance, stall, pending stall, prefetch degree, prefetch distance, and prefetch offset. The effectiveness of hardware data prefetching in reducing cache misses is shown by presenting a square matrix multiplication example. Thereafter the pitfalls of prefetching and possible solutions are discussed.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.5
• /
• pp.306-315
• /
• 2006

In general, the hit ratio of the first level cache is one of the most important factors in determining the performance of computer systems. Prefetching from lower level memory structure is one of the most useful techniques for improving the hit ratio of the first level cache. In this paper, we propose a prefetch on continuous same page access (CSPA) scheme which improves the prefetch efficiency of the instruction cache and reduces prefetch cost at the same time. The proposed CSPA scheme traces the page addresses of executed instructions to count how many times the same memory page is accessed continuously. To increase the prefetch efficiency, the CSPA scheme initiates prefetch only if the number of accesses to the same page exceeds the threshold value. Generally, the size of a L1 cache block is smaller than that of a L2 cache block. Therefore, one L2 cache block contains a number of L1 cache blocks. To reduce the number of unnecessary accesses to the L2 cache due to prefetch, the CSPA scheme enables prefetch only when the missed L1 block and the prefetch L1 block are in the same L2 cache block, leading to reduced prefetch cost. According to our simulations, the proposed prefetching scheme improves the performance by up to 6.7%.

• Journal of KIISE:Computer Systems and Theory
• /
• v.34 no.5_6
• /
• pp.238-248
• /
• 2007

As the demand of employing various PC software on mobile embedded devices which have limited storages has been increased, software streaming services are needed. However it takes too much time to launch software on them because it is transferred through wireless networks. To address this problem, prefetch algorithms are needed. We examined 'Last successor (LS)' algorithm and PPM-based prefetch algorithm as prefetch algorithms for software streaming services. We present 'Probability-base Multi Successor (PMS)' algorithm which is contrived through analyzing evaluations of previous algorithms and characteristics of software streaming services. While LS has one successor per each block, PMS has N successors based on probability which is calculated by PPM-based prefetch algorithm. The hit rate of PMS is similar to that of PPM-base prefetch algorithm and the space overhead is similar to that of LS. We can get good efficiency at the point of memory usage when PMS is applied to software streaming services.

• Journal of Internet Computing and Services
• /
• v.4 no.4
• /
• pp.65-74
• /
• 2003

Recently, It is required that multimedia data display on not only PC but also the Mobile device such as cellar phone, PDA. Mobile divieces has not sufficient of storage space or main memory and has slow network speed. So we can't process the multimedia data with the existing way on PC. This paper presents various methods for MultimediaPresentation which are able to display on the mobile device with the low-bandwidth and small storage. There Is an analysis that Is playable on the mobile device using presentation area devision method and replayable analysis method. If the display of multimedia presentation is impossible from this analysis, to apply prefetch techniques a using EPOB apoint and as User QoS of multimedia data through prefetch techniques have keep, do so that playable. Then, it is proposed to playable smooth multimedia presentation as that each of the presentations through prefetch Scheduling techniques point and quantity of data which are prefetch.

• Journal of KIISE:Computer Systems and Theory
• /
• v.29 no.5
• /
• pp.265-273
• /
• 2002

With the explosive growth of WWW traffic, there is an increasing demand for the high performance Web servers to provide a stable Web service to users. The cluster-based Web server is a solution to core with the heavy access from users, easily scaling the server according to the loads. In the cluster-based Web sewer, a back-end node may not be able to serve some HTTP requests directly because it does not have the requested contents in its main memory. In this case, the back-end node has to retrieve the requested contents from its local disk or other back-end nodes in the cluster. To reduce service latency, we introduce a new prefetch scheme. The back-end nodes predict the next HTTP requests and prefetch the contents of predicted requests before the next requests arrive. We develop three prefetch algorithms bated on some useful information gathered from many clients'HTTP requests. Through trace-driven simulation, the service latency of the prefetch scheme is reduced by 10 ~ 25% as compared with no prefetch scheme. Among the proposed prefetch algorithms, Time and Access Probability-based Prefetch (TAP2) algorithm, which uses the access probability and the inter-reference time of Web object, shows the best performance.

• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.11_12
• /
• pp.608-616
• /
• 2005

A cache architecture is proposed here which evokes prefetch at level 1 cache miss. Existing structures only prefetch at level 2 cache miss. In the proposed cache architecture, level 1 cache miss would select demand fetch block and prefetch block from the level 2 cache and store to level 1 cache and prefetch cache, respectively. According to an experimental analysis using 11 benchmark programs, the hierarchical cache architecture that employs both a level 1 cache prefetcher and a level 2 cache prefetcher obtained a maximum $19\%$ increased performance when compared to the cache architecture that employs only a level 2 cache prefetcher.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.11
• /
• pp.836-852
• /
• 2006

This paper proposes an L1 cache prefetch scheme using an excessively aggressive hardware prefetcher and a hardware prefetch filter having a small direct-mapped filtering cache. A quantitative analysis method has been introduced and applied to analyze nonideal effects of aggressive cache prefetching. From those analysis results, the structure and algorithm of a prefetch filter has been derived and simulated, and the overall system performance has been measured using a cycle-by-cycle cache simulator. Experimental results show that the proposed scheme improves the overall system performance by 18% on the average over several benchmarks