• 제목/요약/키워드: Hyper performance

검색결과 247건 처리시간 0.023초

동시 다중 쓰레딩 마이크로프로세서를 위한 스케줄링 알고리즘의 성능 평가 (Performance Evaluation of a New Scheduling Algorithm for the Simultaneous MultiThreading Microprocessor)

  • 이정훈;김진석
    • 정보처리학회논문지A
    • /
    • 제12A권2호
    • /
    • pp.145-150
    • /
    • 2005
  • 최근 많은 프로세서 제작업체들이 프로세서의 효율을 높이기 위한 방법으로 독립적인 쓰레드들을 한 프로세서 사이클에 동시에 실행시킬 수 있는 동시다중 쓰레딩 기술을 구현하고 있으며 그 예의 하나가 하이퍼쓰레딩이다. 물리 프로세서 안에 여러 개의 논리 프로세서를 가질 수있는 하이퍼쓰레딩 기술은 기존의 여러 개의 독립적인 프로세서들을 갖춘 멀티 프로세싱 환경과는 차이가 있으며, 하이퍼쓰레딩 환경에 알맞은 특정한 작업 할당 방법이 필요하다. 따라서, 본 논문에서는 하이퍼쓰레딩 기술에 적합한 스케줄링 알고리즘을 제안하고 그 성능을 다양한 방법으로 측정해 봄으로서 하이퍼쓰레딩 시스템을 올바르게 인식하고 적절하게 관리하여 효율적인 성능을 기대할 수 있게 되었다.

Grant-Aware Scheduling Algorithm for VOQ-Based Input-Buffered Packet Switches

  • Han, Kyeong-Eun;Song, Jongtae;Kim, Dae-Ub;Youn, JiWook;Park, Chansung;Kim, Kwangjoon
    • ETRI Journal
    • /
    • 제40권3호
    • /
    • pp.337-346
    • /
    • 2018
  • In this paper, we propose a grant-aware (GA) scheduling algorithm that can provide higher throughput and lower latency than a conventional dual round-robin matching (DRRM) method. In our proposed GA algorithm, when an output receives requests from different inputs, the output not only sends a grant to the selected input, but also sends a grant indicator to all the other inputs to share the grant information. This allows the inputs to skip the granted outputs in their input arbiters in the next iteration. Simulation results using OPNET show that the proposed algorithm provides a maximum 3% higher throughput with approximately 31% less queuing delay than DRRM.

Hyper-parameter Optimization for Monte Carlo Tree Search using Self-play

  • Lee, Jin-Seon;Oh, Il-Seok
    • 스마트미디어저널
    • /
    • 제9권4호
    • /
    • pp.36-43
    • /
    • 2020
  • The Monte Carlo tree search (MCTS) is a popular method for implementing an intelligent game program. It has several hyper-parameters that require an optimization for showing the best performance. Due to the stochastic nature of the MCTS, the hyper-parameter optimization is difficult to solve. This paper uses the self-playing capability of the MCTS-based game program for optimizing the hyper-parameters. It seeks a winner path over the hyper-parameter space while performing the self-play. The top-q longest winners in the winner path compete for the final winner. The experiment using the 15-15-5 game (Omok in Korean name) showed a promising result.

HyperDB - A High Performance Data Analysis System Based on Grid Computing Technology

  • Kim, Tae-Kyung;Na, Jong-Hwa;Chon, Wan-Sup
    • Journal of the Korean Data and Information Science Society
    • /
    • 제18권1호
    • /
    • pp.161-174
    • /
    • 2007
  • In this paper, we propose a high performance database cluster system called HyperDB to process OLAP queries efficiently. HyperDB is a virtual database system running on top of internet-connected PCs; the PCs are used for their own purpose at ordinary times, but they are able to participate in the database cluster system at non-office hours. We propose fully logical replication technique and optimal parallel intra-query routing technique for extensibility and performance. Experiment for TPC-R benchmark shows significant performance upgrade compared with conventional approaches.

  • PDF

CFD 기반의 비선형 초탄성 재료의 구조 설계 (The Structural Design for Nonlinear Hyperelastic Materials Based on CFD)

  • 정대석;김지영;이종문;박영철
    • 대한기계학회논문집A
    • /
    • 제30권4호
    • /
    • pp.379-386
    • /
    • 2006
  • The hyper-elastic material has been used gradually and its range was extended all over the industry. The performance prediction of hyper-elastic material was required not only experimental methods but also numerical methods. In this study, we presented the process how to use numerical method for hyper-elastic material and applied it to seat-ring of butterfly valve. The finite element analysis was executed to evaluate the mechanical characteristics of hyper-elastic material. And the optimum model considered conditions and features. According to that model, the load conditions were obtained by using CFD analysis.

Clipping Distortion Suppression of Directly Modulated Multi-IF-over-Fiber Mobile Fronthaul Links Using Shunt Diode Predistorter

  • Han, Changyo;Cho, Seung-Hyun;Sung, Minkyu;Chung, Hwan Seok;Lee, Jong Hyun
    • ETRI Journal
    • /
    • 제38권2호
    • /
    • pp.227-234
    • /
    • 2016
  • Herein, we demonstrate clipping distortion suppression of directly modulated multi-IF-over-fiber links using a simple shunt diode predistorter. The dynamic range of a directly modulated analog fiber optic link is limited by nonlinear distortions caused by laser-diode clipping. We investigate the link performance in the context of carrie-to-noise and distortion ratio (CNDR) and error vector magnitude (EVM) requirements when supporting LTE-A services. We also design an analog predistorter with a shunt-diode structure, and demonstrate experimentally that the predistorter has the ability to suppress clipping-induced third-order intermodulation distortions of the link by at most 14 dB. It also improves the CNDR and EVM of the 4-IF-multiplexed LTE-A carriers by 7 dB and 2.9%, respectively.

S/W Based Frame-Level Synchronization for Irregular Screen Processing System

  • Shin, IlHong;Lee, Seonghee;Lee, Eunjun;Lee, Nam Kyung;Lee, Hyunwoo
    • ETRI Journal
    • /
    • 제38권5호
    • /
    • pp.868-878
    • /
    • 2016
  • A multi-screen environment provides a new opportunity for digital signage applications, where various irregularly shaped screens are employed as an effective means for advertisements and information. A formation method applicable to numerous irregularly shaped screens is proposed to provide a new perspective on digital signage applications. The crucial part of the proposed method comes from the geometric sensing of each screen and the formation of a virtual screen where geometrically aligned content extraction and encoding are employed for content transmission to each screen. In addition, a software-based synchronization method for the proposed system is proposed to address the frame-level synchronization between screens. The experimental results of the proposed method show an improved performance of the frame-level synchronization, where the inconsistency between frames is not identified.

In Search of Corporate Growth and Scaleup: What Strategies Drive Unicorns and Hyper-Growing Companies?

  • Lee, Young-Dall;Oh, Soyoung
    • 한국벤처창업학회:학술대회논문집
    • /
    • 한국벤처창업학회 2021년도 춘계학술대회
    • /
    • pp.33-42
    • /
    • 2021
  • Based on the findings of Lee et al.(2020) and Lee & Oh(2021), this paper aims to fill the gap in our knowledge regarding the relationship between strategic choices and corporate growth by utilizing a novel dataset of 'Unicorn' and 'Hyper-growing' companies. Two previous studies provide coherent findings that the relationship between firms' strategies and their performance should be explored under a more comprehensive framework with consideration of both internal and external factors. Therefore, in this study, we apply a single conceptual framework to two different datasets, which considers the strategy factors as independent variables, and the industry(market) and the firm age as moderating variables. For our dependent variables, valuations for unicorn companies and revenue CAGR for hyper-growing companies are used after categorizing them into three uniform groups. The strategy variables include 'Generic (Cost-leadership, Differentiation, focus) strategies', 'Growth(Organic, M&A) strategies', 'Leading(Pioneer, Fast-follower) strategies', 'Target market(B2B, B2C, B2G, C2C) strategies', 'Global(Global, Local) strategies', 'Digital(Online, Offline) strategies.' For industry(market) factors, it consists of historical growth rate for industries and economic, demographic, and regulatory aspects of states and countries. To overcome the differences in their units, they are also uniformly categorized into multiple groups. Before we conduct a regression analysis, we analyze the industry distribution of the 'Unicorn' and the 'Hyper-growing' companies with descriptive statistics at the integrated and individual levels. Next, we employ hierarchical regression models on Study A('Unicorn' companies in 2019) and Study B('Hyper-growing' companies in 2019) under the same comprehensive framework. We then analyze the relationship between the 'strategy' and the 'performance' factors with two different approaches: 1) an integrated regression model with both the sample of Study A and B and 2) respective regression models on Study A and B. This empirical study aims to provide a complete understanding and a reference to which strategy factors should be considered to promote firms' scale-up and growth.

  • PDF