Soojin Hwang is a postdoctoral researcher in Computer Architecture and System Lab (CASYS) of KAIST, supervised by Prof. Jaehyuk Huh.
Hardware Accelerators
Heterogeneous Systems
[ICCD] Sanghyeon Lee, Hongbeen Kim, Soojin Hwang, Guseul Heo, Minwoo Noh, Jaehyuk Huh, "Throughput-Oriented LLM Inference via KV-Activation Hybrid Caching with a Single GPU", The 43rd IEEE International Conference on Computer Design, 2025
[ISCA] Daehyeon Baek, Soojin Hwang, Jaehyuk Huh, "pSyncPIM: Partially Synchronous Execution of Sparse Matrix Operations for All-bank PIM Architectures", The 51st International Symposium on Computer Architecture, 2024
[ISMM] Taekyung Heo, Seunghyo Kang, Sanghyeon Lee, Soojin Hwang, Joongun Park, Jaehyuk Huh, "Supporting Trusted Virtual Machines with Hardware-based Secure Remote Memory", 2024 ACM SIGPLAN International Symposium on Memory Management, 2024
[TACO] Soojin Hwang, Daehyeon Baek, Jongse Park, Jaehyuk Huh, "Cerberus: Triple Mode Acceleration of Sparse Matrix and Vector Multiplication", ACM Transactions on Architecture and Code Optimization, vol 21 (2), 2024
[IISWC] Soojin Hwang*, Sunho Lee*, Jungwoo Kim, Hongbeen Kim, Jaehyun Huh, "mNPUsim: Evaluating the Effect of Sharing Resources with Multi-Core NPUs", The 2023 IEEE International Symposium on Workload Characterization, 2023 (* Equal Contribution)
[ICCD] Bokyeong Kim, Soojin Hwang, Sanghoon Cha, Chang Hyun Park, Jongse Park and Jaehyuk Huh, "Supporting Dynamic Translation Granularity for Hybrid Memory Systems", The 40th IEEE International Conference on Computer Design, 2022
[PACT] Daehyeon Baek, Soojin Hwang, Taekyung Heo, Daehoon Kim and Jaehyuk Huh, "InnerSP: A Memory Efficient Sparse Matrix Multiplication Accelerator with Locality-aware Inner Product Processing", The 30th International Conference on Parallel Architectures and Compilation Techniques, 2021
Ph.D. in School of Computing
2021.03 - 2025.02
Korea Advance Institute of Science and Technology (KAIST), Daejeon, Republic of Korea
Dissertation Title: Memory-centric Accelerator Systems for Large Sparse Machine Learning Workloads
M.S. in School of Computing
2019.03 - 2021.02
Korea Advance Institute of Science and Technology (KAIST), Daejeon, Republic of Korea
B.S. in School of Computing
2015.03 - 2019.02
Korea Advance Institute of Science and Technology (KAIST), Daejeon, Republic of Korea
Programming Languages: C/C++, Python, Chisel (Scala), CUDA
Simulators: mNPUsim, ZSim, DRAMsim3
Library/Frameworks: Intel MKL, NVIDIA cuBLAS, NVIDIA cuSPARSE, TensorFlow
Tools: Xilinx Vivado, Unity, Blender
2023 Fall: CS610 Parallel Computing TA (Prof. Jaehyuk Huh)
2023 Spring: CS311 Computer Organization TA (Prof. Jaehyuk Huh)
2021 Fall: CS230 System Programming TA (Prof. Jaehyuk Huh)
2021 Spring: CS311 Computer Organization TA (Prof. Jaehyuk Huh)
2019 Fall: CS230 System Programming TA (Prof. Jaehyuk Huh)
2019 Spring: CS492 Parallel Computing TA (Prof. Jaehyuk Huh)
2016 Spring: CS101 Introduction to Programming Tutor (KAIST School of Freshman)
Excellent Teaching Assistant Award at KAIST SoC: CS610 Parallel Computing, Fall 2023
Excellent Teaching Assistant Award at KAIST SoC: CS230 System Programming, Fall 2021
Cum Laude at KAIST, 2019