Publications

Notable Publications

(A full publication list is attached after the notable publication list below)

  • MAESTRO: A Data-Centric Approach to Understand Reuse, Performance, and Hardware Cost of DNN Dataflows
    Hyoukjun Kwon, Prasanth Chatarasi, Michael Pellauer, Angshuman Parashar, Vivek Sarkar, and Tushar Krishna
    IEEE MICRO – Top Picks in Computer Architecture Conferences in 2019 (Top Picks)
    2020
    [Paper]
  • Understanding Reuse, Performance, and Hardware Cost of DNN Dataflows: A Data-Centric Approach
    Hyoukjun Kwon, Prasanth Chatarasi, Michael Pellauer, Angshuman Parashar, Vivek Sarkar, and Tushar Krishna
    In Proc. of the IEEE/ACM International Symposium on Microarchitecture (MICRO)
    Selected as IEEE MICRO Top Picks in Computer Architecture Conferences in 2019
    Final list at student research competition (SRC) at MICRO 2018
    Oct. 2019
    Acceptance Rate: 23.0%
    [Paper][Slides][Project Home Page]
  • MAERI: Enabling Flexible Dataflow Mapping over DNN Accelerators via Reconfigurable Interconnects
    Hyoukjun Kwon, Ananda Samajdar, and Tushar Krishna
    In Proc. of the 23rd ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)
    Honorable Mention : IEEE MICRO Top Picks in Computer Architecture Conferences in 2018
    Mar. 2018
    Acceptance Rate: 17.5%
    [Paper]
  • [Book] Data Orchestration in Deep Learning Accelerators
    Tushar Krishna, Hyoukjun Kwon, Angshuman Parashar, Michael Pellauer, and Ananda Samajdar (Alphabetical order by the last name)
    Synthesis Lectures on Computer Architecture, Morgan & Clay, 2020
    [Book]

Full Publication List

  • C: Conference paper
  • W: Workshop paper
  • J: Journal paper

* Acceptance rates: Added if the statistics were shared.

2025

[C26] Efficient Depth Estimation for Unstable Stereo Camera Systems on AR Glasses
Yongfan Liu, Hyoukjun Kwon
In Proc. of the 2025 Conference on Computer Vision and Pattern Recognition (CVPR)
Jun. 2025
Acceptance Rate: 22.1%
[paper]

[C25] Understanding the Performance Horizon of the Latest ML Workloads with NonGEMM Workloads
Rachid Karami, Sheng-Chun Kao, Hyoukjun Kwon
In Proc. of the 2025 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)
Best paper award
May. 2025
Acceptance Rate: 28.3%
[paper]

[C24] Constrained Dataflow Accelerator for Real-Time Multi-Task Multi-Model Machine Learning Workloads
Jamin Seo, Jianming Tong, Tushar Krishna, Hyoukjun Kwon
In Proc. of the 2025 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)
May. 2025
Acceptance Rate: 28.3%
[paper (to be uploaded)]

[C23] Performance Implications of Multi-Chiplet Neural Processing Units on Autonomous Driving Perception
Mohanad Odema, Luke Chen, Hyoukjun Kwon, Mohammad Al Faruque
In Proc. of the Design, Automation and Test in Europe Conference (DATE)
Mar. 2025
[paper (to be uploaded)]

2024

[C22] SCAR: Scheduling Multi-Model AI Workloads on Heterogeneous Multi-Chiplet Module Accelerators
Mohanad Odema, Luke Chen, Hyoukjun Kwon, Mohammad Al Faruque
In Proc. of the IEEE/ACM International Symposium on Microarchitecture (MICRO)
Nov. 2024
Acceptance Rate: 22.7%
[paper]

[C21] Characterizing the Accuracy – Efficiency Trade-off of Low-rank Decomposition in Language Models
Chakshu Moar, Faraz Tahmasebi, Michael Pellauer, Hyoukjun Kwon
In Proc. of the 2024 IEEE International Symposium on Workload Characterization (IISWC)
Sep. 2024
[paper]

[C20] DREAM: A Dynamic Scheduler for Dynamic Real-time Multi-model ML Workloads
Seah Kim, Hyoukjun Kwon, Jinook Song, Jihyuck Jo, Yu-Hsin Chen, Liangzhen Lai, Vikas Chandra
In Proc. of the 28thACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)
Mar. 2023 (Presented at ASPLOS 2024)
Acceptance Rate: 25.2%
[paper]

2023

[C19] XRBench: An Extended Reality (XR) Machine Learning Benchmark Suite for the Metaverse
Hyoukjun Kwon, Krishnakumar Nair, *Jamin Seo, *Jason Yik, Debabrata Mohapatra, Dongyuan Zhan, Jinook Song, Peter Capak, Peizhao Zhang, Peter Vajda, Colby Banbury, Mark Mazumder, Liangzhen Lai, Ashish Sirasao, Tushar Krishna, Harshit Khaitan, Vikas Chandra, Vijay Janapa Reddi (*: equal contribution)
Sixth Conference on Machine Learning and Systems (MLSys)
Jun. 2023
Acceptance Rate: 22%
[paper][project homepage][poster]

2022

[W2] MetaBench: Real-Time Multi-Modal Benchmark for Metaverse
Hyoukjun Kwon, Krishnakumar Nair, Jinook Song, Colby Banbury, Mark Mazumder, Peter Capak, Yu-Hsin Chen, Liangzhen Lai, Tushar Krishna, Harshit Khaitan, Vikas Chandra,Vijay Janapa Reddi
The third Workshop on Benchmarking Machine Learning Workloads on Emerging Hardware (MLBench)
Received the best paper award at the workshop
Sep. 2022

[C18] A Formalism of DNN Accelerator Flexibility
Sheng-Chun Kao, Hyoukjun Kwon, Michael Pellauer, Angshuman Parashar, Tushar Krishna
In Proc of the 2022 ACM SIGMETRICS/IFIP PERFORMANCE Joint International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS)
Jun. 2022
[paper]

[C17] Multi-scale High-resolution Vision Transformer for Semantic Segmentation
Jiaqi Gu, Hyoukjun Kwon, Dilin Wang, Wei Ye, Yu-Hsin Chen, Liangzhen Lai, Vikas Chandra, and David Z. Pan
In Proc of the 2022 Conference on Computer Vision and Pattern Recognition (CVPR)
Jun. 2022
[paper]

2021

[J6] Marvel: A Data-Centric Approach for Mapping Deep Learning Operators on Spatial Accelerators
Prasanth Chatarasi, Hyoukjun Kwon, Angshuman Parashar, Michael Pellauer, Tushar Krishna and Vivek Sarkar
ACM Transactions on Architecture and Code Optimization (TACO)
2021

[J5] Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication
Gordon E Moon, Hyoukjun Kwon, Geonhwa Jeong, Prasanth Chatarasi, Sivasankaran Rajamanickam, Tushar Krishna
IEEE Transactions on Parallel and Distributed Systems (TPDS)
2021

[C16] Extending Sparse Tensor Accelerators to Support Multiple Compression Formats
Eric Qin, Geonhwa Jeong, William Won, Sheng-Chun Kao, Hyoukjun Kwon, Sudarshan Srinivasan, Dipankar Das, Gordon E. Moon, Sivasankaran Rajamanickam and Tushar Krishna
In Proc of the 35th IEEE International Parallel & Distributed Processing Symposium (IPDPS)
May. 2021

[C15]  Heterogeneous Dataflow Accelerators for Multi-DNN Workloads
Hyoukjun Kwon , Liangzhen Lai, Michael Pellauer, Tushar Krishna, Yu-Hsin Chen, and Vikas Chandra
In Proc. of the 27th IEEE International Symposium on High-Performance Computer Architecture (HPCA)
Mar. 2021
Acceptance Rate: 24.4% (63/258)
[Paper]

[C14]  Dataflow-Architecture Co-Design for 2.5D DNN Accelerators using Wireless Network-on-Package
Robert Guirado and Hyoukjun Kwon (equal contribution), Sergi Abadal, Eduard Alarcon, and Tushar Krishna
In Proc. of the 26th Asia and South Pacific Design Automation Conference (ASP-DAC)
Jan. 2021
Acceptance Rate: 34.2% (111/327)

2020

[J4] Flexion: A Quantitative Metric for Flexibility in DNN Accelerators
Hyoukjun Kwon, Michael Pellauer, Angshuman Parashar, and Tushar Krishna
IEEE Computer Architecture Letters (CAL)
2020

[J3] Architecture, Chip, and Package Co-design Flow for 2.5D Integration of Reusable IP Chiplets
Jinwoo Kim, Gauthaman Murali, Heechun Park, Eric Qin, Hyoukjun Kwon, Venkata Chaitanya Krishna, Nihar Dasari, Arvind Singh, Minah Lee, Hakki Torun, Kallol Roy, Madhavan Swaminathan, Saibal Mukhopadhyay, Tushar Krishna, and Sung Kyu Lim
IEEE Transactions on Very Large Scale Integration (VLSI) Systems (VLSI)
2020
[Paper]

[B1] Data Orchestration in Deep Learning Accelerators
Tushar Krishna, Hyoukjun Kwon, Angshuman Parashar, Michael Pellauer, and Ananda Samajdar
Synthesis Lectures on Computer Architecture, Morgan & Clay, 2020
[Book]

[C13] Co-Exploration of Neural Architectures and Heterogeneous ASIC Accelerator Designs Targeting Multiple Tasks
Lei Yang, Zheyu Yan, Meng Li, Hyoukjun Kwon, Liangzhen Lai, Tushar Krishna, Vikas Chandra, Weiwen Jiang, Yiyu Shi
In Proc. of the 57th Annual Design Automation Conference (DAC)
Jul. 2020
[Paper]

[J2] MAESTRO: A Data-Centric Approach to Understand Reuse, Performance, and Hardware Cost of DNN Dataflows
Hyoukjun Kwon, Prasanth Chatarasi, Michael Pellauer, Angshuman Parashar, Vivek Sarkar, and Tushar Krishna
IEEE MICRO – Top Picks in Computer Architecture Conferences in 2019 (Top Picks)
2020
[Paper]

[C12] SIGMA: A Sparse and Irregular GEMM Accelerator with Flexible Interconnects for DNN Training
Eric Qin, Ananda Samajdar, Hyoukjun Kwon, Vineet Nadella, Sudarshan Srinivasan, Dipankar Das, Bharat Kaul, and Tushar Krishna
In Proc. of The 26th IEEE International Symposium on High-Performance Computer Architecture (HPCA)
Best paper award
Feb. 2020
[Paper]

2019

[C11] Understanding the Impact of On-Chip Communication on DNN Accelerator Performance
Robert Guirado, Hyoukjun Kwon, Sergi Abadal, Eduard Alarcon, and Tushar Krishna
In Proc. of the 26th IEEE International Conference on Electronics Circuits and System (ICECS)
Nov. 2019

[C10] Understanding Reuse, Performance, and Hardware Cost of DNN
Dataflows: A Data-Centric Approach
Hyoukjun Kwon, Prasanth Chatarasi, Michael Pellauer, Angshuman Parashar, Vivek Sarkar, and Tushar Krishna
In Proc. of the IEEE/ACM International Symposium on Microarchitecture (MICRO)
Selected as IEEE MICRO Top Picks in Computer Architecture Conferences in 2019
Final list at student research competition (SRC) at MICRO 2018
Oct. 2019
[Paper][Slides][Project Home Page]
Acceptance Rate: 23.0% (79/344)

[C9] Architecture, Chip, and Package Co-design Flow for 2.5D Integration of Reusable IP Chiplets
Jinwoo Kim, Gauthaman Murali, Heechun Park, Eric Qin, Hyoukjun Kwon, Venkata Chaitanya Krishna, Nihar Dasari, Arvind Singh, Minah Lee, Hakki Torun, Kallol Roy, Madhavan Swaminathan, Saibal Mukhopadhyay, Tushar Krishna, and Sung Kyu Lim
In Proc. of the 56th Annual Design Automation Conference (DAC)
Jun. 2019
[Paper]

[C8] mRNA: Enabling Efficient Mapping Space Exploration on a Reconfigurable Neural Accelerator
Zhongyuan Zhao, Hyoukjun Kwon, Sachit Kuhar, Weiguang Sheng , Zhigang Mao, and Tushar Krishna
In Proc. of the IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)
Mar. 2019
[Paper]
Acceptance Rate: 29.5% (26/88)

2018

[J1] A Communication-driven Approach for Designing Flexible DNN Accelerators
Hyoukjun Kwon, Ananda Samajdar, and Tushar Krishna
IEEE Micro Special Issue on Hardware Acceleration (IEEE Micro)
Nov./Dec. 2018
[Paper]

[C6] MAERI: Enabling Flexible Dataflow Mapping over DNN Accelerators via Reconfigurable Interconnects
Hyoukjun Kwon, Ananda Samajdar, and Tushar Krishna
In Proc. of the 23rd ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS)
Honorable Mention in IEEE MICRO Top Picks 2019
Mar. 2018
Acceptance Rate: 17.5% (56/319)
[Paper]

[C5] MAERI: Enabling Flexible Dataflow Mapping over DNN Accelerators via Reconfigurable Interconnects
Hyoukjun Kwon, Ananda Samajdar, and Tushar Krishna
In Inaugural SysML Conference (SysML; now MLSys) (not-archived)
Feb. 2018
[Paper]

[W1] Spoofing Prevention via RF Power Profiling in Wireless Network-on-Chip
Brian Lebiednik, Sergi Abadal, Hyoukjun Kwon, and Tushar Krishna
In Proc. of the 3rd International Workshop on Advanced Interconnect Solutions and Technologies for Emerging Computing Systems (AISTECS)
Jan. 2018
[Paper]

2017

[C4] Rethinking NoCs for Spatial Neural Network Accelerators
Hyoukjun Kwon, Ananda Samajdar, and Tushar Krishna
In Proc. of the ACM International Symposium on Network-on-Chip (NOCS)
Oct. 2017
[Paper][Code]

[C3] Adaptive Manycore Architectures for Big Data Computing
Janardhan Rao Doppa, Ryan Gary Kim, Mihailo Isakov, Michel A. Kinsy, Hyoukjun Kwon, and Tushar Krishna
In Proc. of the ACM International Symposium on Network-on-Chip (NOCS), (special session paper)
Oct. 2017
[Paper]


[C2] Proving Flow Security of Sequential Logic via Automatically-Synthesized Relational Invariants
Hyoukjun Kwon, William Harris, and Hadi Esmaeilzadeh
In Proc. of the IEEE Computer Security Foundations Symposium (CSF)
Aug. 2017
Acceptance Rate: 34% (32/94)
[Paper]

[C1] OpenSMART: Single-Cycle Multi-hop NoC Generator in BSV and Chisel
Hyoukjun Kwon and Tushar Krishna
In Proc. of the IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)
Apr. 2017
Acceptance Rate: 29% (24/81)
[Paper]