• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.10b
• /
• pp.1541-1544
• /
• 2001

Java virtual machine의 성능을 향상시키기 위해 "JIT(Just-in-Time)"code generator가 고안되었다[3], JIT code generator는 스택기반의 자바 바이트 코드를 레지스터 기반의 native machine code로 변환해 주는 역할을 수행하여 바이트 코드의 번역시간을 줄여준다. 그러나 JIT 는 많은 레지스터의 사용을 야기시키므로 효율적인 레지스터 allocation 정책이 필요하고 스택과 레지스터 간의 traffic 을 가중시킨다. 그러므로 본 논문에서는 자바 바이트 코드의 효율적인 stack allocation 정책을 JIT code generator에 적용함으로 레지스터와의 traffic을 줄이는 방법을 제시하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06a
• /
• pp.570-573
• /
• 2011

Dalvik 가상 머신은 최적화가 쉽지 않았던 스택 기반의 가상 머신인 자바 가상 머신(JVM)과 달리 레지스터 기반의 가상 머신이므로 여러 최적화 기법을 적용할 수 있다. 따라서 Dalvik 가상 머신을 위한 새로운 레지스터 할당기법이 필요하다. 본 논문에서는 dx tool을 거친 레지스터 기반의 Dalvik byte code를 레지스터 재 할당하여 최적화하고, 이를 Dalvik JIT으로 보내어 다시 한번 레지스터 재할당 함으로써 최적화 효율을 높일 수 있는 최적화 기법을 제안하였다. 또한 제안 기법과 Dalvik JIT의 복잡도를 비교함으로써 제안기법을 검증하였다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.35 no.12
• /
• pp.540-551
• /
• 2008

Today's microprocessor designs are not free from temperature as well as power consumption. As processor technology scales down, an on-chip circuitry increases power density, which incurs excessive temperature (hotspot) problem. To tackle thermal problems cost-effectively, Dynamic Thermal Management (DTM) has been suggested: DTM techniques have benefits of thermal reliability and cooling cost. However, they require trade-off between thermal control and performance loss. This paper proposes a dual integer register file structure to minimize the performance degradation due to DTM invocations. In on-chip thermal control, the most important functional unit is an integer register file. It is the hotspot unit because of frequent read and write data accesses. The proposed dual integer register file migrates read data accesses by adding an extra register file, thus reduces per-unit dynamic power dissipation. As a result, the proposed structure completely eliminates localized hotspots in the integer register file, resulting in much less performance degradation by average 13.35% (maximum 18%) improvement compared to the conventional DTM architecture.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.41 no.10
• /
• pp.85-94
• /
• 2004

In this paper, a register allocation techlique that translates memory accesses to register accesses Is presented to enhance embedded software performance. In the proposed method, a source code is profiled to generate a memory trace. From the profiling results, target functions with high dynamic call counts are selected, and the proposed register allocation technique is applied only to the target functions to save the compilation time. The memory trace of the target functions is searched for the memory accesses that result in cycle count reduction when replaced by register accesses, and they are translated to register accesses by modifying the intermediate code and allocating Promotion registers. The experiments where the performance is measured in terms of the cycle count on MediaBench and DSPstone benchmark programs show that the proposed method increases the performance by 14% and 18% on the average for ARM and MCORE, respectively.

• Journal of IKEEE
• /
• v.3 no.1 s.4
• /
• pp.150-160
• /
• 1999

This paper presents a new register transfer level synthesis method to overcome the disadvantages of the previous register transfer level synthesis systems. The previous register transfer level synthesis systems first translate from a hardware description language to sequential circuits inadequately. Secondly, the systems separate registers and combinational circuits and then optimize only combinational circuits. This paper describes their disadvantages and then proposes a new method to overcome their shortcomings. This paper also shows the effectiveness of the proposed method by using the proposed method at designing the controller of a surveillance system and the 8-bit signed multiplier.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.1 s.307
• /
• pp.1-8
• /
• 2006

This thesis proposes Arithmetic Shift Register(ASR) which can be used as pseudo random number generator. Arithmetic Shift. Register is defined as progression that multiplies random number D , not 0 or 1 at initial value which is not 0, and it is represented as ASR-D in this thesis. Irreducible polynomial that t which makes $'D^k=1'$ satisfies uniquely as $'t=2^n-1'$ over. $GF(2^n)$ is the characteristic polynomial of ASR-D , and the cycle of Arithmetic Shift Register has maximum cycle as $'2^n-1'$. Galois Linear Feedback Shift Register corresponds to ASR-2-1. Therefore, Arithmetic Shift Register proposed in this thesis generalizes Galois Linear Feedback Shift Register. Linear complexity of ASR-D over$GF(2^n)$ is $'n{\leq}LC{\leq}\frac{n^2+n}{2}'$ and in comparison with existing Linear Feedback Shift Register stability is high. The Software embodiment of arithmetic shift register proposed in this thesis is efficient than that of existing Linear Shift Register and hardware complexity is equal. Arithmetic shift register proposed in this thesis can be used widely in various fields such as cipher, error correcting codes, Monte Carlo integral, and data communication etc along with existing linear shift register.

• Proceedings of the KIEE Conference
• /
• 2006.10d
• /
• pp.98-100
• /
• 2006

본 논문은 기존의 고속 인쇄기용 레지스터 컨트롤러가 고가의 외국 제품을 사용해 온 것에 반해 이를 대체 할 뿐 아니라 저렴한 가격의 레지스터 칸트롤러를 개발 하는 것을 목표로 하고 기존 250mpm(meter per minute)의 두 배인 500mpm의 고속 인쇄에서도 사용할 수 있도록 레지스터 컨트롤러를 개발해 오차 보정을 좀 더 정확하고 신속하게 하는 것에 그 목적이 있다.

• Journal of KIISE:Software and Applications
• /
• v.28 no.2
• /
• pp.201-209
• /
• 2001

고성능 마이크로 프로세서들은 성능 향상을 위해 ILP를 지원한다. 병렬성을 극대화시키기 위해서는 많은 성능 저해 요인들을 제거해야 한다. 최근에는 컴파일러의 역할을 증대시켜 이러한 요인들을 줄이기 위한 노력들이 활발히 진행되고 있다. 본 논문에서는 성능 저해 요인인 조건 분기 처리를 위하여 조건 실행과 레지스터 할당을 결합함으로써 메모리로의 대피를 최소화하고 병렬성을 향상시킬 수 있는 개선된 레지스터 할당 알고리즘을 제안한다. 제안한 방법을 적용하여 실험한 결과 간섭 그래프의 에지수가 4.47% 감소되었고 그 결과 요구되는 대피 변수의 수도 21.35% 감소되었다. 그리고 기존의 방법에 비해 19.38%의 성능 향상 결과를 얻었다. 결국 본 레지스터 할당 기법은 조건 실행을 통해 조건 분기 명령을 제거하여 기본 블록 내의 명령어 수를 증가시켜 병렬처리의 기회를 증진시키고 조건 분석을 통해 간섭 그래프의 불필요한 에너지를 제거시켜 보다 효율적인 레지스터 할당을 실현함으로써 제안한 방법의 타당성을 검증하였다.

• Journal of the Korea Society of Computer and Information
• /
• v.5 no.3
• /
• pp.1-8
• /
• 2000

Prepass code scheduling(code scheduling before register allocation), the register lifetimes may be lengthened, which may increase the amount of data dependence relations. So, it makes difficult to allocate the registers because of complex interference graph. In this paper, to improve that defect, propose an 2-phase coloring method. At first phase-1 assign the registers to variables which have long live ranges. Secondly, phase-2 allocate the registers to remained variables to minimize the register allocation cost. Experimental results shown that proposed method is more efficient scheme than Chaitin's scheme when prepass code scheduling.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.11 no.2
• /
• pp.161-169
• /
• 2011

This paper presents a new register transfer level synthesis methodology for efficient SOC system design. The previous register transfer level synthesis systems first translate from a hardware description language to sequential circuits inadequately. Secondly, the systems separate registers and combinational circuits and then optimize only combinational circuits. This paper describes their disadvantages and then proposes a new method to overcome their shortcomings. This paper also shows the effectiveness of the proposed method by using the proposed method at designing the controller of a surveillance system.