Ian

Minsoo Rhu

Senior Research Scientist @ NVIDIA

Google Scholar | Linkedin | Email

About me

I am a researcher in the areas of computer architecture with particular interests in the intersection of high-performance computing (HPC) systems and deep learning. I received my Ph.D. in computer engineering from the University of Texas at Austin in 2014 (thesis advisor: Mattan Erez). I received my M.S. and B.S. degrees from KAIST (Korea Advanced Institute of Science and Technology) and Sogang University, in 2009 and 2007, respectively. My research experience spans several areas in computing including ASIC designs, computer system architecture, and application and workload characterizations with an emphasis on deep neural networks.

Publications

MICRO-49 vDNN: Virtualized Deep Neural Networks for Scalable, Memory-Efficient Neural Network Design
Minsoo Rhu, Natalia Gimelshein, Jason Clemons, Arslan Zulfiqar, and Stephen W. Keckler
The 49th IEEE/ACM International Symposium on Microarchitecture (MICRO-49), Taipei, Taiwan, Oct. 2016
[PDF] [Slide] [Code]

HPCA-23 Architecting an Energy-Efficient DRAM System For GPUs
Niladrish Chatterjee, Mike O'Connor, Donghyuk Lee, Daniel R. Johnson, Stephen W. Keckler, Minsoo Rhu, and William J. Dally
The 23rd IEEE International Symposium on High Performance Computer Architecture (HPCA-23), Austin, TX, Feb. 2017
[PDF]

MICRO-48 CLEAN-ECC: High Reliability ECC for Adaptive Granularity Memory System
Seong-Lyong Gong, Minsoo Rhu, Jungrae Kim, Jinsuk Chung, and Mattan Erez
The 48th IEEE/ACM International Symposium on Microarchitecture (MICRO-48), Waikiki, HI, Dec. 2015
[PDF]

HPCA-21 Priority-Based Cache Allocation for Throughput Processors
Dong Li*, Minsoo Rhu, Daniel R. Johnson, Mike O'Connor, Mattan Erez, Doug Burger, Donald S. Fussell, and Stephen W. Keckler
The 21st IEEE International Symposium on High Performance Computer Architecture (HPCA-21), San Francisco, CA, Feb. 2015
*Dong Li is a co-first author of this work.
[PDF] [Slide]

ISLPED'14 GPUVolt: Characterizing and Mitigating Voltage Noise in GPUs
Jingwen Leng, Yazhou Zu, Minsoo Rhu, Meeta Sharma Gupta, and Vijay Janapa Reddi
The IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED-2014), La Jolla, CA, Aug. 2014
[PDF]

MICRO-46 A Locality-Aware Memory Hierarchy for Energy-Efficient GPU Architectures
Minsoo Rhu, Michael Sullivan, Jingwen Leng, and Mattan Erez
The 46th IEEE/ACM International Symposium on Microarchitecture (MICRO-46), Davis, CA, Dec. 2013
[PDF] [Slide]

ISCA-40 Maximizing SIMD Resource Utilization in GPGPUs with SIMD Lane Permutation
Minsoo Rhu and Mattan Erez
The 40th IEEE/ACM International Symposium on Computer Architecture (ISCA-40), Tel-Aviv, Israel, Jun. 2013
[PDF] [Slide]

HPCA-19 The Dual-Path Execution Model for Efficient GPU Control Flow
Minsoo Rhu and Mattan Erez
The 19th IEEE International Symposium on High Performance Computer Architecture (HPCA-19), Shenzhen, China, Feb. 2013
[PDF] [Slide]

ISCA-39 CAPRI: Prediction of Compaction-Adequacy for Handling Control-Divergence in GPGPU Architectures
Minsoo Rhu and Mattan Erez
The 39th IEEE/ACM International Symposium on Computer Architecture (ISCA-39), Portland, OR, Jun. 2012
[PDF] [Slide]

TCSVT Optimization of Arithmetic Coding for JPEG2000
Minsoo Rhu and In-Cheol Park
IEEE Transactions on Circuits and System for Video Technology, Vol.20, No.3, pp.446-451, Mar. 2010
[PDF]

ICIP-2009 Memory-less Bit-Plane Coder Architecture for JPEG2000 with Concurrent Column-Stripe Coding
Minsoo Rhu and In-Cheol Park
IEEE International Conference on Image Processing 2009 (ICIP 2009), Cairo, Egypt, Nov. 2009
[PDF]

ICIP-2009 Architecture Design of a High-Performance Dual-Symbol Binary Arithmetic Coder for JPEG2000
Minsoo Rhu and In-Cheol Park
IEEE International Conference on Image Processing 2009 (ICIP 2009), Cairo, Egypt, Nov. 2009
[PDF]

SiPS-2009 A Novel Trace-Pipelined Binary Arithmetic Coder Architecture for JPEG2000
Minsoo Rhu and In-Cheol Park
IEEE Workshop on Signal Processing Systems 2009 (SiPS 2009), Tampere, Finland, Oct. 2009
[PDF]

Patents

9,477,526 Cache utilization and eviction based on allocated priority tokens (issued 10/25/16)