• Journal of Information Processing Systems
• /
• 제17권1호
• /
• pp.51-62
• /
• 2021

On-chip caches of graphics processing units (GPUs) have contributed to improved GPU performance by reducing long memory access latency. However, cache efficiency remains low despite the facts that recent GPUs have considerably mitigated the bottleneck problem of L1 data cache. Although the cache miss rate is a reasonable metric for cache efficiency, it is not necessarily proportional to GPU performance. In this study, we introduce a second key determinant to overcome the problem of predicting the performance gains from L1 data cache based on the assumption that miss rate only is not accurate. The proposed technique estimates the benefits of the cache by measuring the balance between cache efficiency and throughput. The throughput of the cache is predicted based on the warp occupancy information in the warp pool. Then, the warp occupancy is used for a second bypass phase when workloads show an ambiguous miss rate. In our proposed architecture, the L1 data cache is turned off for a long period when the warp occupancy is not high. Our two-level bypassing technique can be applied to recent GPU models and improves the performance by 6% on average compared to the architecture without bypassing. Moreover, it outperforms the conventional bottleneck-based bypassing techniques.

• 국제초고층학회논문집
• /
• 제5권2호
• /
• pp.95-103
• /
• 2016

One of the key problems in the design of high-rise commercial complex is how to guide reasonable pedestrian distribution in commercial space. In this study, pedestrian distribution in three high-rise commercial complexes in Shanghai and Hong Kong was studied using spatial configuration analysis software Space Syntax and quantification of physical elements in commercial spaces, such as functional attractiveness, entrances, escalators, level variations and passage width. Additionally, in an attempt to integrate functions with spatial integration and spatial depth, two combination variables, the spatial coefficient of function (IF) and spatial depth coefficient of function (F/D), were proposed. The results of the correlation analysis and multiple regression analyses reflected the following: (1) Regarding the influence on pedestrian distribution, there was a synergistic and complementary relationship between function and space; (2) The comprehensive flow distribution analytic model could successfully interpret flow distribution in high-rise commercial complexes and its R Square ranged up to about 70% in the three cases; (3) The spatial coefficient of function (IF) and spatial depth coefficient (F/D) could effectively integrate functions and spatial configuration, which could help close the gap between over-emphasis on function in commercial research and the lack of consideration of function in space-syntax analysis.

• 산업공학
• /
• 제13권3호
• /
• pp.528-538
• /
• 2000

HLA is a general-purpose software architecture for distributed simulation designed to support a wide range of simulation approaches and application. The US DoD's HLA for modeling and simulation can certainly be regarded as the state of the art in distributed simulation. It is a mandatory standard for military simulation. The purpose of this paper is to describe applications of HLA/RTI in multiple domains across the manufacturing systems society. In many and large scale industrial systems, enormous data is generated, and is to be managed in an effective way. It needs a high performance common network library. Furthermore, it must satisfy the real function of system facilities as much as possible. The RTI is an implementation of the interface specification, provided as a set of services. Some applications focusing on the area of a small manufacturing system were demonstrated. The integration could be achieved using the HLA, together with interface modules for each of the subsystems. We have found that HLA/RTI are cable of meeting the functional requirements for a given system environment.

• 대한전자공학회논문지
• /
• 제27권8호
• /
• pp.1201-1210
• /
• 1990

This paper proposes on-chip instruction and data cache memories on RISC reduced instruction set computer) architecture which supports fast instruction fetch and data read/write, and enables RISC processor under research to obtain high performance. In the execution of HLL(high level language) programs, heavily used local scalar variables are stored in large register file, but arrays, structures, and global scalar variables are difficult for compiler to allocate registers. These problems can be solved by on-chip Instruction/Data cache. And each cycle of instruction fetch, pad delay causes the lowering of the processors's performance. Cache memories are designed in CMOS technology and SRAM(static-RAM), that saves layout area and power dissipation, is used for instruction and data storage. To speed up and support RISC processor's piplined architecture efficiently, hardwired logic technology is used overall circuits i cache blocks. The schematic capture and timing simulation of proposed cache memorises are performed on Apollo DN4000 workstation using Mentor Graphics CAD tools.

• Asia pacific journal of information systems
• /
• 제5권1호
• /
• pp.52-85
• /
• 1995

This paper develops a new approach to dealing with qualitative reasoning processes involved in managerial decisions, drawing upon choice strategies that have been developed within the general framework of multi-criteria decision making. Issues such as choices under uncertainty and preference formulation are addressed. An MCDM DSS intended to assist in high-level management decisions must focus on helping the decision maker to properly define the problem by providing a structure to it and to dynamically evaluate the alternative courses of action. A conceptual architecture is developed and presented to propose a general model for designing decision support systems specifically designed to assist in MCDM in a managerial context. A commercial loan approval judgment case is described to illustrate the real-world situation where decisions are made under fuzziness and usually require a high degree of intuition and subjective judgment. Development of a prototype system intended to partially represent application of the architecture is described.

• International Journal of Naval Architecture and Ocean Engineering
• /
• 제5권3호
• /
• pp.392-403
• /
• 2013

For the analysis of radiating noise problems in medium-to-high frequency ranges, the Energy Flow Boundary Element Method (EFBEM) was developed. EFBEM is the analysis technique that applies the Boundary Element Method (BEM) to Energy Flow Analysis (EFA). The fundamental solutions representing spherical wave property for radiating noise problems in open field and considering the free surface effect in underwater are developed. Also the directivity factor is developed to express wave's directivity patterns in medium-to-high frequency ranges. Indirect EFBEM by using fundamental solutions and fictitious source was applied to open field and underwater noise problems successfully. Through numerical applications, the acoustic energy density distributions due to vibration of a simple plate model and a sphere model were compared with those of commercial code, and the comparison showed good agreement in the level and pattern of the energy density distributions.

• 국제초고층학회논문집
• /
• 제6권4호
• /
• pp.333-343
• /
• 2017

Across the world, building energy codes are becoming stricter, demanding higher levels of energy performance with each issuance. Some locations have taken initiatives to eliminate operational emissions altogether by requiring buildings to be carbon neutral. However, while the objectives of carbon neutrality are without doubt statement worthy, we believe that once operational performance has been tackled to a reasonable level of performance the sights should be trained on a different objective-life; cycle carbon. This paper defines what we mean by life cycle carbon neutrality and presents an approach toward reducing it.

• 한국조경학회지
• /
• 제23권1호
• /
• pp.23-38
• /
• 1995

ELEECTRE is a less demanding and potentially Bffective multiattribute evaluation method for pre-editing proeedure. The method, however, was developed based on probabilistic mathematicsand its reliability has never'been emprieally tested. This article examined the reliability of ELECTRo under various conditions of linear and non-linear utility func- tions using a Monte-Carlo shnulation, and found the method eausea a considerably high rate of error. To enhance the reliability of ELECTRE, two ways of modi(ication - adjust- ment of threshold values of concordance and discordance matrices, and adjustment of the required nunlber of probably-dominant alternatives to eliminate a probably-dominated alternative - were suggested, and their effctiveness was also tesed by a Monte-Carlo simulation. The simulation result chows that these modifications cponsiderably improve the reliability of ELECTRE, and yet maintain a reasonably high level of efficiency. Through these modifications, ELECTRE can be used as an efficient and feliable pre-editing method. Ways to implement ELECTRE in the construction of plainning decision support systens were discussed. The impli(els combination of this method and other multiattribute evaluation methods will help to create more effective decision suport systems.

• 한국조경학회지
• /
• 제42권2호
• /
• pp.19-32
• /
• 2014

본 연구는 도시공원의 관리운영을 위한 민간기금 조성측면에서 부산시 APEC기념공원을 대상으로 가상평가법(Contingent Valuation Method: CVM)을 이용하여, 쾌적한 도시공원 유지를 위한 거주민이 인식하는 경제적 가치를 평가했다. 그 결과, 부산시 APEC기념공원의 관리운영을 위해 월 가구당 지불 가능한 추정금액은 5,673원에서 8,358원으로 평가되었다. 추정금액을 공원 유치권역의 가구로 확대하여 살펴보면, 2012년 편성된 공원 관리예산의 2.2~3.3배 수준으로 분석되어, 지역 거주민들이 인식하는 쾌적한 공원 유지를 위해 지불의사금액이 편성된 관리예산보다 높은 것으로 나타났다. 그리고 도시공원의 관리운영을 위한 지불의사금액에 영향을 주는 응답자의 사회적 요인으로는 연령, 교육수준, 소득으로 분석되었으며, 연령대가 높고, 교육수준이 높은 응답자에서 지불의사가 높은 것으로 평가되었다. 또한, 거주민의 공원이용 및 인식에 따라 도시공원 관리운영을 위한 지불의사금액에 많은 영향을 미치고 있는 것으로 나타났다. 특히 공원이용 빈도가 높은 거주민이 지불의사가 높으며, 주1회 이상 공원이용자가 전체 지불의사금액보다 1.1배 높은 추정금액으로 평가되었다. 그리고 공원의 필요성을 인식하는 것보다 공원을 만족하는 것이 더 높은 지불의사금액으로 이어지는 것을 밝혀냈다. 따라서 공원 만족도 제고를 위한 공원시설관리 및 이용프로그램 운영이 필요하며, 공원에 대한 인식을 기부로 연결할 수 있는 공원문화 보급정책이 필요하다고 판단된다.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제31권8호
• /
• pp.1073-1082
• /
• 2004

본 논문에서는 변화하는 환경에서 에이전트의 인지 정보로부터 움직이는 물체의 운동모델을 미리 알 수 없는 경우에도 적용할 수 있는 적응적인 행동을 생성하는 방법을 제안한다. 전통적인 에이전트의 지능제어 방법은 환경에 대해 알고 있는 정보를 이용한다는 제약 때문에 강건하지만 다양하고 복잡한 환경에 적용할 수 얼었다. 환경에 대한 정보가 없는 상황에서 에이전트가 자율적으로 행동하기 위해서는 행동 기반의 방법이 적합하며, 실제와 같은 변화는 환경에서 에이전트의 적응적 행동을 위해서는 상황을 미리 추론하고 대처하는 능력이 필요하다. 움직이는 장애물 피하기는 변화하는 환경에서의 적응적 행동생성의 가능성을 보여줄 수 있는 문제이기 때문에 다양한 방법으로 연구되고 있다. 본 논문에서는 고정된 장애물뿐만 아니라 움직이는 장애물을 인지하고 피하는 적응적인 행동을 생성하기 위한 2단계의 제어 구조를 제안한다. 1단계는 상황을 인지하고 자율적으로 행동을 생성하는 행동 네트워크 구조이고 2단계는 변화하는 상황을 추론하고 제어정보를 1단계로 전달하는 베이지안 네트워크 구조이다. 시뮬레이터를 이용한 실험을 통해 제안한 방법으로 고정된 장애물과 움직이는 장애물을 피하고 목적지를 찾아가는 것을 확인할 수 있었다.