• Journal of Electrical Engineering and information Science
• /
• v.1 no.2
• /
• pp.15-26
• /
• 1996

MPAs(Massively Parallel Architectures) should address two fundamental issues for scalability: synchronization and communication latency. Dataflow architecture faces problems of excessive synchronization overhead and inefficient execution of sequential programs while they offer the ability to exploit massive parallelism inherent in programs. In contrast, MPAs based on von Neumann computational model may suffer from inefficient synchronization mechanism and communication latency. DAVRID (DAtaflow/Von Neumann RISC hybrID) is a massively parallel multithreaded architecture which takes advantages of von Neumann and dataflow models. It has good single thread performance as well as tolerates synchronization and communication latency. In this paper, we describe the DAVRID architecture in detail and evaluate its performance through simulation runs over several benchmarks.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.43 no.2 s.344
• /
• pp.51-59
• /
• 2006

Von Neumann architecture suffers from the interface between the central processing unit and the memory, which is called 'Von Neumann bottleneck' In this paper, we propose a scalable general-purpose associative processor (AP) based on content-addressable memory (CAM) which solves this problem and is suitable for the search-oriented applications. We propose an efficient instruction set and a structural scalability to extend for larger applications. We define twelve instructions and provide some reduced instructions to speed up which execute two instructions in a single instruction cycle. The proposed AP performs in a bit-serial, word-parallel fashion and can be considered as a 32-bit general-purpose parallel processor with a massively parallel SIMD structure. We design and simulate a maximum/minumum search greater-than/less-than search, and parallel addition to verify the proposed architecture. The algorithms are executed in a constant time O(k) regardless of the number of input data.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.10 no.4
• /
• pp.175-183
• /
• 1985

Dataflow computers exhibit a high degree of parallelism which can not be obtained easily with the conventional von-Neumann architecture. Since many instructions are ready for execution simultaneously, concurrency can be easily achieved by the multiple processors modified the dataflow machine. This paper describes a FFT Butterfly algorithm for dataflow computation and evaluates the performance by the speed up factor of that algorithm through the simulation approach by the time-accelation method.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.10a
• /
• pp.134-137
• /
• 2004

지능 홈-케어 시스템 또는 외부 통신 채널의 환경 인식이 가능한 모바일 통신기기와 같은 상황 인식 시스템이 외부 상태를 감지하여 현재 상창을 인식하고 대처하기 위해서는 수 백개 이상의 규칙들을 이용한 추론을 필요로 한다. 이들 규칙들의 효과적인 추론을 위해서는 룰-베이스 시스템에 기반을 둔 추론 기법을 적용시킬 수 있다 이 룰-베이스 시스템의 추론 규칙의 매칭을 위해서 RETE 알고리즘이 사용되어 왔다. 하지만 RETE 알고리즘은 그 특성상 Von Neumann 구조의 컴퓨터 시스템에서는 규칙의 증가에 따른 그 성능의 저하가 필연적이다. 본 논문에서는 RETE 네트워크를 이용한 추론을 효과적으로 수행할 수 있는 RETE 네트워크 하드웨어 가속기의 구조에 대해서 논한다. 이 RETE 네트워크 하드웨어 가속기은 Von Neumann의 구조적 제약점을 병렬처리 구조를 사용하여 제거하였다.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.1
• /
• pp.167-180
• /
• 1996

In the scalar processing oriented machine scalar operations must be performed for the vector processing as many as the number of vector components. So called a vector processing mechanism by the von Neumann operational principle. Accessing vector data hasto beperformed by theevery pointing ofthe instruction or by the address calculation of the ALU, because there is only a program counter(PC) for the sequential counting of the instructions as a memory accessing device. It should be here proposed that an access unit dimensionally to address components has to be designed for the compensation of the organizational hardware defect of the conventional concept. The necessity for the vector structuring has to be implemented in the instruction set and be performed in the mid of the accessing data memory overlapped externally to the data processing unit at the same time.

• Electronics and Telecommunications Trends
• /
• v.33 no.5
• /
• pp.121-134
• /
• 2018

The Von Neumann based architecture of the modern computer has dominated the computing industry for the past 50 years, sparking the digital revolution and propelling us into today's information age. Recent research focus and market trends have shown significant effort toward the advancement and application of artificial intelligence technologies. Although artificial intelligence has been studied for decades since the Turing machine was first introduced, the field has recently emerged into the spotlight thanks to remarkable milestones such as AlexNet-CNN and Alpha-Go, whose neural-network based deep learning methods have achieved a ground-breaking performance superior to existing recognition, classification, and decision algorithms. Unprecedented results in a wide variety of applications (drones, autonomous driving, robots, stock markets, computer vision, voice, and so on) have signaled the beginning of a golden age for artificial intelligence after 40 years of relative dormancy. Algorithmic research continues to progress at a breath-taking pace as evidenced by the rate of new neural networks being announced. However, traditional Von Neumann based architectures have proven to be inadequate in terms of computation power, and inherently inefficient in their processing of vastly parallel computations, which is a characteristic of deep neural networks. Consequently, global conglomerates such as Intel, Huawei, and Google, as well as large domestic corporations and fabless companies are developing dedicated semiconductor chips customized for artificial intelligence computations. The AI Processor Research Laboratory at ETRI is focusing on the research and development of super low-power AI processor chips. In this article, we present the current trends in computation platform, parallel processing, AI processor, and super-threaded AI processor research being conducted at ETRI.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.2
• /
• pp.291-298
• /
• 2023

SRAM-based in-memory computing is one of the technologies to solve the bottleneck of von Neumann architecture. In order to achieve SRAM-based in-memory computing, it is essential to design efficient SRAM bit-cell. In this paper, we propose a low-power differential sensing 8+T SRAM bit-cell which reduces power consumption and improves circuit performance. The proposed 8+T SRAM bit-cell is applied to ripple carry adder which performs SRAM read and bitwise operations simultaneously and executes each logic operation in parallel. Compared to the previous work, the designed 8+T SRAM-based ripple carry adder is reduced power consumption by 11.53%, but increased propagation delay time by 6.36%. Also, this adder is reduced power-delay-product (PDP) by 5.90% and increased energy-delay- product (EDP) by 0.08%. The proposed circuit was designed using TSMC 65nm CMOS process, and its feasibility was verified through SPECTRE simulation.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.23 no.2
• /
• pp.225-235
• /
• 1986

Dataflow computers exhibit a high degree of parallelism which can not be obtained easily with the conventional von-Neumann architecture. Since many instructions are ready for execution simultaneously, concurrency can easily by achieved by the multiple processors modified the data-flow machine. In paper, we describe an improved dataflow architecture which is designed by adding the demand propagation network to the MIT dataflow machine. and show the improved performance by the execution time and the efficiency of processing elements through simulation with the time acceleration method.

• Proceedings of the KIEE Conference
• /
• 1987.07b
• /
• pp.965-968
• /
• 1987

Conventional major electronic switching systems based on stored program control employ a Von Neumann styled control processor. It has strict limitations such that it essentially lacks concurrency in executing instructions, which have brought the software bottleneck problem, and the capabilities of call processing are restricted by expanding system's capacity. In this paper, a new architecture of call control processor based on the data flow system is proposed, aiming at fundamental resolution for these limitations. The processor has a number of advantages in such as expansibility of system's capacity, parallel processing of calls, and so on.

• Electronics and Telecommunications Trends
• /
• v.35 no.4
• /
• pp.34-41
• /
• 2020

The existing Von Neumann architecture places limits to data processing in AI, a booming technology. To address this issue, research is being conducted on computing architectures and artificial neural networks that simulate neurons and synapses, which are the hardware of the human brain. With high-speed, high-throughput data communication infrastructures, photonic solutions today are a mature industrial reality. In particular, due to the recent outstanding achievements of artificial neural networks, there is considerable interest in improving their speed and energy efficiency by exploiting photonic-based neuromorphic hardware instead of electronic-based hardware. This paper covers recent photonic neuromorphic studies and a classification of existing solutions (categorized into multilayer perceptrons, convolutional neural networks, spiking neural networks, and reservoir computing).