이론적인 선인출 모델

김성백,김한일 濟州大學校工科大學産業技術硏究所 1996 尖端技術硏究所論文集 Vol.7 No.2

Prefetching aims at reducing memory latency by fetching, in advance, data that are likely to be requested by the processor in a near future. The effectiveness of prefetching is determined by how accurate the prediction on the data that will be requested by the processor. Most previous studies on prefetching were limited to proposing a particular prefetch scheme and its performance evaluation. paying little attention to theoretical aspects of prefetching. This paper focuses on the theoretical aspects of instruction prefetching.For this purpose,we propose an ideal prefetching model that makes use of perfect history information. Based on this theoretical model, we analyzed upper limits on the prefetching accuracies of the SPEC benchmarks. The results show that the prefetching accuracy is very high when there is no cache. However. as the size of the instruction cache increases. the prefetching accuracy drops drastically. For example, in the case of the spice benchmark, the prefetching accuracy drops from 53% to 39% when the cache size increases from 2Kbyte to 16Kbyte (assuming 16byte block size). These results indicate that as the cache size increases, most localities are captured by the cache and that instruction prefetching based on the information extracted from the references that missed in the cache suffers from prediction inaccuracies.

製紙슬럿지 소각재의 燒結處理와 溫度의 相關性에 관한 硏究

김환기,이성백,정흥렬 全北大學校 附設 都市및環境硏究所 1995 都市 및 環境硏究 Vol.10 No.1

Paper and pulps industry is improving with development of paper and publication culture, also disposal of paper sludge and flyash has induced significant factor of development in the paper industry because of shortage in disposal area and inhabitant demonstration. There fore this studies persuit fundamentle database of flyash specification in the physical and chemical analysis for recycling architectural materials, etc. 1. This study investigates production of papersludge and flyash chemicalcomposition, physical and chemical specification ( ignition loss, specific weight, sintering temperture. And density etc.) 2. Especially focuses in the characteristics of the sintering treatment and sintering temperature of flyash. 3. Accomplishes in cut of cost through save energy & reuse of waste materials. Environmentally conservation of our nature has to be carried by ourselves.

주정폐수의 혐기성 소화처리시와 포기처리시 처리특성에 관한 연구

金煥起,李城白,黃譚勇 全北大學校 附設 都市및環境硏究所 1996 都市 및 環境硏究 Vol.11 No.1

The wastewater of alcohol manufactory plant has very high suspended soild and BOD concentration, acidity, and high temperture. Thus most of the plants drive on anaerobic digestion method of biological treatment because of using a lot of generated methane gas and possible to disinfect pathogenic bacteria in sluge and wastewater. But because of the problem in the concentration of anaerobic treatment water, second treatment is necessary 2-step activated treatment for increasing treatment efficiency. The purpose of this study was performed to know a difference between last 1-step digestion 1-step aeration treatment and 2-step digestion 2-step aeration treatment with retention time in digestion and aerator. The obtained results can be summarized as follows; 1. The concentration characteristic of alchohol production wastewater was analyzed COD 50,000∼80,000㎎/ℓ, BOD 22,000∼41,000㎎/ℓ, SS 20,000∼45,000㎎/ℓ, pH 4.1∼4.2, and 75∼80℃ according to the raw material of alchohol. 2. In case of the anaerobic digestion method of rice-barley wastewater, average COD removal rate of 2-step digestion appeared to 91.9% compared to average COD removal rate 59.3% on 1-step digestion. It is necessary to drive above 20 day of retention time in metane degestor because of increasing organic load causing the decrease of pH and the increase of VFA/COD within aerator. 3. Alcohol wastewater of tapioka is generally much treatment efficiency of 2-step digestion than that of 1-step digestion, but below 2.7㎏-COD/㎥·d of COD bulk load showed the possibility of treatment at 1-step and 2-step digestion method. 4. Amount of gas generation in anaerobic digestion was almost similar at 1-step and 2-step digestion. 5. In the result of aerobic treatment, treatment efficiency of 2-step consecutive aeration was highly about 3% compared to 1-step combined aeration and in the condition of COD 4,000㎎/ℓ, COD load 800㎏ -COD/d, COD volume-load 0.8㎏-COD/㎥·d were difficult two treatment methods both. 6. pH and VFA concentrations are considered as important factors to determine VFA restriction and the harmfulness of fatty acid was seemed to be related to pH. VFA in 1-step digestion appeared strong change with the increase of input wastewater quantity, but 2-step digestion was treated to stable change in output COD 1,230∼2,246㎎/ℓ.

모래여과에 의한 하수처리 유출수의 인제거 특성에 관한 연구

김환기,이성백,신은섭 全北大學校 附設 都市및環境硏究所 1998 都市 및 環境硏究 Vol.13 No.1

Phosphate remaining in wastewater treatment must be removed for the conservation of water quality in a water system. The concentration of phosphate can be decreased from a physic-chemical method for the discharge water of a sewage disposal plant. In this study, it was performed to examine the adsorptive characteristics of sand filter through batch and column experiments in saturated soil. Batch experiment evaluates the concentration change of phosphate with retention time and column experiment performs to analyze a breakthrough curve for phosphate adsorption. And the batch experiment result of adsorptive equilibrium was fitted for the Langmuir model and the Freundlich model. The results obtained from this study were summarized as follows; 1. In the batch experiment, the adsorptive capacity of PO_4^-3 by sand was below 47% of adsorption rate for retention 24 hours in the range of initial concentration 2.0㎎/ℓ ∼4.6㎎/ℓ. 2. The Langmuir and the Freundlich models on the adsorptive equilibrium were fitted well to both in a good correlation coefficient. 3. In the column experiment, the adsorptive breakthrough curve for input PO_4^-3(15.0㎎/ℓ) occurred to about 30ℓ. And output concentration was almost decided at initial breakthrough curve.

TDMA 구현 이더네트에 기반한 실시간 RPC 시스템의 설계

이정훈,이봉규,김성백 濟州大學校 基礎科學硏究所 1997 基礎科學硏究 Vol.10 No.1

본 논문은 요청-확인-응답 시맨틱을 갖는 실시간 원격 프로시쥬어 호출을 효율적으로 지원할 수 있는 통신 시스템을 설계하고 그 성능을 평가한다. 이 통신 시스템은 시분할 다중 접근 방식이 구현된 이더네트를 기반으로 하고 있으며 원격 프로시쥬어 호출에 관련된 메시지들을 그들의 특성에 따라 스케쥴하여 서버의 스케쥴과 쉽게 동기화시킨다. 서버와 통신 시스템 스케쥴 동기화에 의해 서버의 수행시간을 효율적으로 사용할 수 있으므로 실시간 원격 프로시쥬어 호출의 종료시한 만족도를 향상시킬 수 있다. 요청의 도착시간, 서비스 시간 및 여유 시간 등 다양한 환경 인자 설정에 의한 모의 실험에 의해 제안된 통신 시스템이 기존의 토큰 버스나 시분할 다중 접근 방식보다 실시간 원격 프로시쥬어 호출의 실시간 성능을 향상시킴을 보인다. This paper proposes and analyzes a communication mechanism capable of supporting real-time RPC(Remote Procedure Call) which is based on request-acknowledgment-reply semantic. As an enhanced protocol of TDMA-implemented Ethernet, the proposed CS(Communication System) dynamically schedules RPC-related messages according to their attributes, making a network schedule be easily combined with a server schedule. As RPC schedule produced by combining server schedule and communication schedule can obviate the waste computation time at the server, more real-time RPC requests can meet their timing constraints. Simulation result shows that the RPC over the proposed CS outperforms those over token bus and TDMA on the various network parameters, such as RPC arrival time, service time and slack.

IT Fusion-based Approaches to Education: Tackling the Problems of Global Education, Constructing Mastery Learning Model, and Creating Competitive-type Learning Games

Seong Baeg Kim 세계환경사회거버넌스학회 2017 World Environment and Island Studies Vol.7 No.2

The Analysis of Prefetch Accuracy Using the Ideal Instruction Prefetch Model

Kim, Seong-Baeg,Lee, Jung-hoon 濟州大學校 師範大學 科學敎育硏究所 1999 科學敎育 Vol.16 No.-

Prefetching aim at reducing memory latency by fetching, in advance, instructions and data that are likely to be requested by the processor in a near future. The effectiveness of prefetching depends on how accurate the prediction on the needed instructions and data is Most previous studies on prefetching were limited to proposing a particular prefetch scheme and evaluating its performance, giving little attention to its theoretical aspects This paper focuses on the limit of instruction prefetching. For this purpose, we propose an ideal model that makes use of a post-morten analysis of a program execution to derive an upper bound on prefetch accuracy. Based on this model, we analyzed the upper bounds on prefetch accuracy (i.e., theoretically achievable prefetch accuracy) of programs from the SPEC benchmarks The results show that when there is no instruction cache the upper bounds are very high. However, as the size of the instruction cache increases, the prefetch accuracy bound drops drastically. For example. in the case of the spice benchmark, the achievable prefetch accuracy drops from 53 % to 39 % when the cache size increases from 2 Kbytes to 16 Kbytes (assuming 16 byte block size). One implication of this result is that when the cache size is over implication of ths some threshold and the prefetch miss penalty is high, instruction prefetching may degrade the overall performance rather than improve it.

흰쥐의 실험적 심근경색증에서 Captopril 및 Propranolol의 장기간 투여가 좌심실 형태학적 변형에 미치는 효과

김철민 ( Kim Cheol Min ),정욱성 ( Jeong Ug Seong ),백상홍 ( Baeg Sang Hong ),전승석 ( Jeon Seung Seog ),김종상 ( Kim Jong Sang ),김재형 ( Kim Jae Hyeong ),최규보 ( Choe Gyu Bo ),홍순조 ( Hong Sun Jo ) 대한내과학회 1992 대한내과학회지 Vol.42 No.6

저자들은 흰쥐에서 심근경색을 유발시킨 후 7일부터 42일까지 captopril및 propranolol을 투여한 후 나타나는 좌심실의 형태학적 변형을 알아보기 위해 경색의 크기, 경색확대율, 좌심실 반경, 심실중격의 두께 등을 측정하여 다음과 같은 곁과를 얻었다. 1) waterf의 좌심실 반경, 경색확대율, 심내막 둘레는 가상대조군에 비해 현저히 증가하였다. 2) captopril군의 좌심실 반경, 심실중격 두께는 Water군에 비해 현저히 감소하였다. 3) propranolol군의 좌심실 반경은 water군에 비해 감소 경향을 보였으나, 심실 중격의 두께의 차이는 보이지 않았다. 이상의 결과 휜쥐의 실험적 심근 경색증후 captopril의 장기 투여는 좌심실의 형태학적 변형에서 좌심실 화대와 대상성 심실비대를 약화시켰으며, propranolol의 장기투여는 좌심실 확대를 약화시켰으나 대상성 심실비대에는 영향을 주지 못하였다. To determine the effects of long-term therapy with captopril or propranolol on left ventricular remodeling, we studied 101 Sprague-Dawley rats with experimental myocardial infarction. The animals were randomized to recieve captopril (2000 ㎎/L) or propranolol (750 ㎎/L) in their drinking water, or water alone for 42 days. Cross sections of the left ventricle were prepared for morphometric measurements of infarct size, infarct expension rate, left ventricular radius, and septal thickeness. The results were as follows: 1) In untreated rats with experimental myocardial infarction, left ventricular remodeling was revealed (total water group). The left ventricular radius (3.1+0.4㎜) and infarct expantion rate (37.8+6.4%) was larger than those of the sham operated rats (p<0.005, p<0.01). 2) The septal thickness of the total captopril group (1.58+0.20㎜) was smaller than that of the total water group (1.89±0.35㎜)(p<0.05), and the septal thickness of the captopril group with moderate myocardial infarction (1.62±0.13 ㎜) was also smaller than that of the propranolol group with moderate myocardial infarction (1.84±0.13 ㎜)(p<0.05). The left ventricular radius of the captopril group with moderate myocardial infarction (2.6k0.3 mm) was smaller than that of the propranolol group (3.0±0.3 ㎜)(p < 0.05). 3) There was no significant difference in the septal thickness between the total propranolol group and total water group. The left ventricular radius of total pro-pranolol group was smaller than that of water group insignificantly. It was concluded that long-term therapy of captopril had effects in rats with infarct, resulting in significant blunting of left ventricular dilation and septal hypertrophy, and long-term therapy of propranolol had effects, resulting in blunting of left ventricular dilation but no decreasing of septa1 hypertrophy.

명령어 선인출 예측 정확도의 한계에 관한 연구

김성백(Seong Baeg Kim),민상렬(Sang Lyul Min),김종상(Chong Sang Kim) 한국정보과학회 2000 정보과학회논문지 : 시스템 및 이론 Vol.27 No.8

선인출은 프로세서에 의해 사용될 데이타를 예측하여 미리 프로세서 근처에가져오므로써 메모리 지연 시간을 줄이는 기법이다. 선인출의 효율성은 미래에 사용될 데이타를 얼마나 정확하게 예측하는가(선인출 예측 정확도)에 따라 결정된다. 기존의 명령어 선인출에 관한 연구들은 특정 선인출 기법의 제안 및 성능 평가에 그치고 있어서 명령어 선인출의 특성이 체계적으로 분석 정리되지 못하고 있다. 이에 본 논문에서는 명령어 선인출의 예측 정확도에 대해서 이론적으로 분석하여 이의 한계를 알아보고자 한다. 그 방안으로 명령어 선인출 상한 모델이라는 이론적인 선인출 모델을 제안하고 이 모델을 기반으로 명령어 선인출에 대해 체계화된 분석을 한다. 특히 이러한 연구 결과로써 궁극적으로 시스템 성능을 효과적으로 향상시킬 수 있는 효율적인 명령어 선인출을 가능하게 하는 데 그 목적이 있으므로 주로 명령어 선인출 효율성 측면에서 분석을 시도하였다. 이러한 선인출 모델을 이용하여 본 논문에서는 SPEC 벤치마크 프로그램들의 명령어 선인출 예측 정확도의 한계를 이론적으로 분석하였다. 그 결과로 캐쉬가 없는 경우에는 선인출 정확도가 매우 높게 나타남을 보였다. 반면에 캐쉬가 있을 경우에는 캐쉬 크기가 커짐에 따라 선인출의 정확도가 급격히 떨어짐을 관찰하였다. 예를 들어 spice의 경우 블록크기가 16바이트이고 직접사상 캐쉬에서 캐쉬 크기가 2K 바이트와 16K 바이트일 때 이론적으로 가능한 최대 선인출 정확도가 각각 53%, 39%로 크게 떨어지는 것을 관찰하였다. 캐쉬의 크기가 커질수록 선인출로 메모리 지연 시간을 줄일 수 있는 명령어 참조의 많은 부분을 캐쉬가 처리하게 되고 또한 캐쉬에서 접근 실패된 명령어 참조는 그 참조 행태가 불규칙하여 예측이 어렵기 때문에 일정 크기 이상의 명령어 캐쉬를 사용하는 경우 명령어 선인출을 사용하는 것은 전체 시스템 성능의 향상에 큰 도움이 되지 않음을 이론적으로 규명하였다. Prefetching aims at reducing memory latency by fetching, in advance, data that are likely to be requested by the processor in a near future. The effectiveness of prefetching is determined by how accurate the prediction on the needed instructions and data is. Most previous studies on prefetching were limited to proposing a particular prefetch scheme and its performance evaluation, paying little attention to theoretical aspects of prefetching. This paper focuses on the theoretical aspects of instruction prefetching. For this purpose, we propose a clairvoyant prefetch model that makes use of perfect history information. Based on this theoretical model, we analyzed upper limits on the prefetch prediction accuracies of the SPEC benchmarks. The results show that the prefetch prediction accuracy is very high when there is no cache. However, as the size of the instruction cache increases, the prefetch prediction accuracy drops drastically. For example, in the case of the spice benchmark, the prefetch prediction accuracy drops from 53% to 39% when the cache size increases from 2Kbyte to 16Kbyte (assuming 16byte block size). These results indicate that as the cache size increases, most localities are captured by the cache and that instruction prefetching based on the information extracted from the references that missed in the cache suffers from prediction inaccuracies.

포스터 발표 : 신동맥 스텐트 삽입술로 치료한 섬유근성 이형성증에 병발한 자발성 신동맥 박리 1예

백현정 ( Baeg Hyeon Jeong ),김대중 ( Kim Dae Jung ),여호명 ( Yeo Ho Myeong ),임영환 ( Im Yeong Hwan ),김정아 ( Kim Jeong A ),강우헌 ( Kang U Heon ),김범 ( Kim Beom ),허우성 ( Heo U Seong ),김윤구 ( Kim Yun Gu ),오하영 ( O Ha Yeon 대한신장학회 2003 춘계학술대회 초록집 Vol.22 No.1