New Publication (MICRO 2026)
Our paper, Non-Speculative Redundant Load Elimination, has been accepted to MICRO 2026.
YONSEI UNIVERSITY / COMPUTER ARCHITECTURE
Learning
from nature.
Building for
discovery.
We explore new ways to design computer architectures and systems—pursuing performance, efficiency, and ideas that endure.
GAUDI Lab
General-Purpose Architectures
with Unleashed Design Innovations
A NAME. A WAY OF THINKING.
Named in tribute to Antoni Gaudí, our lab aspires to carry his spirit into computing: learning from nature, working with humility, and devoting ourselves to work that takes time.
RESEARCH AREAS
From microarchitecture to complete systems
Our research spans the computing stack, connecting fundamental architecture ideas with emerging platforms and real workloads.
01
CPU Microarchitecture
Speculation, dynamic scheduling, SMT, and heterogeneous multi-core architectures.
Explore → 02GPU & Accelerators
Energy-efficient GPGPU, NPU, virtual memory, and on-chip accelerator architectures.
Explore → 03Memory & CXL
Emerging interconnects, memory expansion, and heterogeneous cooperative computing.
Explore → 04System Orchestration
Resource management across CPUs, accelerators, memory, cache, and I/O devices.
Explore → 05Near-Data Processing
Processing in memory, storage, and networks for data-intensive applications.
Explore →SELECTED IMPACT
Research recognized at leading venues
CPU MicroarchitectureReconstructing Out-of-Order Issue QueueMICRO 2022
GPU ArchitectureWarped-Compaction: Maximizing GPU Register File Bandwidth Utilization via Operand CompactionHPCA 2025
Memory & CXLDemystifying CXL Memory with Genuine CXL-Ready Systems and DevicesMICRO 2023
System ArchitectureA4: Microarchitecture-Aware LLC Management for Datacenter Servers with Emerging I/O DevicesISCA 2025
Near-Data ProcessingUPP: Universal Predicate Pushdown to Smart StorageISCA 2025
LATEST
News from the lab
New Project (IITP)
Our research proposal, Artificial Intelligence Innovation Graduate School, has been accepted.
New Project (KIAT)
Training Program for Semiconductor Core IP Design Specialists has been accepted.
New Project (NRF)
Our project on CPU-NPU contention in ultra-low-latency on-device AI SoCs has been accepted.
JOIN GAUDI LAB
Build the next computing architecture with us.
We welcome highly motivated postdoctoral researchers, Ph.D. candidates, and M.S. students. Scholarships are available for qualified applicants.