http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
동시연산 다중 digit을 이용한 직렬 십진 곱셈기의 설계
유창헌(ChangHun Yu),김진혁(JinHyuk Kim),최상방(SangBang Choi) 대한전자공학회 2015 전자공학회논문지 Vol.52 No.4
본 논문에서는 직렬 십진 곱셈기의 성능을 향상시키는 방안을 제안하고 다중 digit을 동시에 연산하는 방안을 제안한다. 제안하는 직렬 십진 곱셈기는 부분 곱 생성단계의 2배수, 4배수를 생성하기 위한 인코딩 모듈을 없애고 쉬프트 연산만으로 부분 곱을 생성해 지연시간을 감소시킨다. 또한 다중 digit 연산을 이용해 연산의 횟수를 줄인다. 제안하는 직렬 십진 곱셈기의 성능을 평가하기 위해서 Synopsys사의 Design Compiler를 이용하여 SMIC사의 110nm CMOS 공정 라이브러리로 합성하였다. 그 결과 제안한 곱셈기는 기존의 직렬 십진 곱셈기와 비교해 전체 면적은 4% 증가하였지만, 전체 지연시간은 5% 감소함을 보였다. 또한 동시 연산 수가 증가함에 따른 제안한 다중 digit 곱셈기의 면적과 지연시간의 trade-off를 확인하였다. In this paper, the method which improves the performance of a serial decimal multiplier, and the method which operates multiple-digit simultaneously are proposed. The proposed serial decimal multiplier reduces the delay by removing encoding module that generates 2X, 4X multiples, and by generating partial product using shift operation. Also, this multiplier reduces the number of operations using multiple-digit operation. In order to estimate the performance of the proposed multiplier, we synthesized the proposed multiplier with design compiler with SMIC 110nm CMOS library. Synthesis results show that the area of the proposed serial decimal multiplier is increased by 4%, but the delay is reduced by 5% compared to existing serial decimal multiplier. In addition, the trade off between area and latency with respect to the number of concurrent operations in the proposed multiple-digit multiplier is confirmed.
김록범,김병권,김유미,홍영습,유창헌,김대선 환경독성보건학회 2013 환경독성보건학회지 Vol.28 No.-
Objectives The aim of this study was to determine the association between low-level mercuryexposure and neurobehavioral functions in adults living in coastal regions of Korea. Methods We selected 172 adults aged 20-65 years living in a city in the coastal region of Korea. A sociodemographic survey was conducted, mercury levels in the blood, urine, andhair were measured, and the associations according to computerized neurobehavioral testswere determined using univariate analysis. After adjustment for associated variables, a multivariatelinear regression analysis was performed. Results The geometric mean mercury levels in the blood, urine, and hair were 5.41 μg/L(range, 0.00-15.84 μg/L), 1.17 μg/g-creatinine (range, 0.00-32.86 μg/g-creatinine), and 1.37mg/kg (range, 0.42-6.56 mg/kg), respectively. Variables that were associated with simplereaction time according to the neurobehavioral test results were age and urine mercury level. Variables associated with choice reaction time were the recent use of Korean traditionalmedicine and urine mercury level. Variables associated with the right-hand finger tappingspeed test were age, gender, smoking behavior, education level, monthly household income,and urine mercury level. Variables associated with the left-hand finger tapping speedtest were age, gender, education level, and urine mercury level. After adjustment for associatedvariables, there was no significant association between urine mercury level and simplereaction time (β=25.96; p =0.47), choice reaction time (β=50.37; p =0.32), or the numberof left-hand finger taps (β=-1.54; p =0.21). However, urine mercury level was significantlyassociated with the number of right-hand finger taps (β=-3.86; p =0.01). Conclusions We found no evidence that low-level mercury exposure in adults is associatedwith deficits in neurobehavioral functions. A longer follow-up study is required to confirmthis conclusion.