Computer Architecture

Research Goals and Vision

The computer architecture research at the KAIST School of Electrical Engineering aims to design the next generation of computer hardware and software systems and its interface for current and emerging application domains like Artificial Intelligence (AI). This research area focuses on optimizing the organization of computer architecture and systems, ease of programming, and system performance. The advancement in computer architecture has led to the design of next-generation computer systems that enable the execution of entirely new forms of applications. For example, AI technology, which is revolutionizing many fields in the IT industry, has been made practical by AI semiconductor technologies such as Graphics Processing Units (GPUs) and Neural Processing Units (NPUs), the achievement of which are made possible by innovative research results in the field of computer architecture.

Key Research Areas

Processor Design

Energy-efficient Processor : The focus is on designing high-performance, low-power processors that can efficiently handle complex computations. The research emphasizes multi-core and parallel processing techniques, developing advanced compilers and hardware architectures to maximize computational speed and energy efficiency.
Processor Security and Reliability : Processor design and implementations that enhance its security and reliability.

관련교수

FACULTIES

Memory and Storage Systems

Next-Generation Memory/Storage Solution : Researchers work on improving data access speed through advanced memory hierarchies, exploring next-generation memory technologies such as High Bandwidth Memory (HBM) and Non-Volatile Memory (NVM). They also optimize storage system management for efficiency and performance.

관련교수

FACULTIES

Neural Network Accelerators

AI Accelerator Integration : The development of hardware accelerators aims to speed up artificial intelligence (AI) and machine learning tasks. The focus is on leveraging hardware such as GPUs, FPGAs, and ASICs to improve neural network processing speed while minimizing power consumption.

관련교수

FACULTIES

Interconnection Networks

Next-Generation Interconnection Network : Efficient interconnection networks are designed for data transfer in high-performance computing systems. Researchers optimize network bandwidth, latency, and scalability, aiming to enhance the communication architecture and performance in large-scale distributed systems.

관련교수

FACULTIES

Quantum and Neuromorphic Computing

Quantum computing : Quantum computing focuses on developing new computational paradigms capable of solving problems much faster than traditional computers.
AI Accelerator Integration : AI accelerators mimicking the human brain are being explored to develop systems that can efficiently process neural networks.

관련교수

FACULTIES

Recent related activities in Computer Architecture

See below for specifc ongoing research topics related to Computer Architecture of KAIST EE.

Research

Research Areas