Guohui Wang (王国晖)

---

I am a reseach engineer at Snapchat working on computer vision and deep learning. Before joining Snapchat, I was a senior GPGPU engineer in graphics team at Qualcomm. I mainly focused on benchmark development, performance analysis and optimization for GPGPU on mobile platforms. I am interested in GPGPU, mobile computing, computer vision, computer architecture, and signal processing.

I got my Ph.D. degree in electrical and computer engineering from Rice University, Houston, Texas. Before coming to the U.S., I got a M.S. degree in CS from Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China, and a B.S. degree in EE from Peking University, Beijing, China.

Projects

• Efficient computer vision and deep learning for mobile AR applications.
• GPGPU acceleration of machine learning inference engine.
• Mobile GPU compute solutions and performance analysis.
• GPGPU accelerated image processing and computer vision on mobile platforms.
• GPGPU-based accelerator for high performance DSP algorithms.
• Parallel VLSI architecture for wireless communication systems.

Software

• OpenCL-Z Android: OpenCL information utility for Android.
  
  • Detect availability of OpenCL;
  • Display OpenCL-capable platform and device information in details;
  • Measure raw compute power using micro-benchmark;
  • Now available on Google Play:


• ezSIFT: an easy-to-use standalone open-source SIFT library.
  
  • Software implementation of the scale-invariant feature transform (SIFT) algorithm;
  • Written in C/C++;
  • Doesn't require other 3rd-party packages;
  • Download source code: https://github.com/robertwgh/ezSIFT.


• Tool: Android native program launcher.
  
  • A tool allowing developers to launch native program on un-rooted devices without connecting the devices to an ADB shell with USB cable.
  • Download Android Native Program Launcher from https://github.com/robertwgh/AndroidNativeLauncher..


• cuLDPC: CUDA implementation of LDPC decoding algorithm
  
  • This is a highly efficient CUDA implementation of layered LDPC decoding algorithm. Our SASP2011, Asilomar2011, GlobalSIP2013 papers are based on this code. (Please note that the code used in the papers might not be exactly the same; but it should not be hard to modify this code to reproduce the results in the paper. We do believe that there are still a lot of room to further improve the performance given new GPU programming techniques and new hardware features during the past few years.)
  • Download source code: https://github.com/robertwgh/cuLDPC.

Technical Notes

• Tutorial - part 1: Using OpenCV Nonfree Module (SIFT, SURF) in Android NDK Projects
  
  • Demonstrate how to build OpenCV Nonfree feature module for Android and use it in an Android NDK project.
  • Webpage: http://web.guohuiwang.com/technical-notes/sift_surf_opencv_android.


• Tutorial - part 2: Use OpenCV Nonfree module (SIFT, SURF) in Android Applications via JNI
  
  • Demonstrate how to use OpenCV Nonfree feature module in an Android application with JNI.
  • Webpage: http://web.guohuiwang.com/technical-notes/opencv_nonfree_android_jni_demo.


• Mastering Android NDK Build System - Part 1: Techniques with ndk-build
  
  • Advanced techniques of using Android NDK building system.
  • Webpage: http://web.guohuiwang.com/technical-notes/androidndk1.


• Mastering Android NDK Build System - Part 2: Standalone toolchain
  
  • Tutorial and examples of using NDK standalone toolchain.
  • Webpage: http://web.guohuiwang.com/technical-notes/androidndk2.

Publications

[Book chapter]
• Embedded Systems Networking: Applications, case studies, and technologies, Elsevier. In preparation.
• High-Level Design Tools for Complex DSP Applications, DSP for Embedded and Real-Time Systems: Expert Guide, Elsevier, 2012. ISBN-13: 9780123865359. Link at Amazon.com.
  Yang Sun, Guohui Wang, Bei Yin, Joseph R. Cavallaro, and Tai Ly
  

[Talks]
• Massively Parallel Signal Processing for Wireless Communication Systems
  Michael Wu and Guohui Wang
  GPU Technology Conference (GTC) 2013. March 18-21, 2013, San Jose, California.
  [Link to slides and video]

[Journal Papers]
• Parallel Interleaver Design for a High Throughput HSPA+/LTE Multi-Standard Turbo Decoder
  Guohui Wang, Hao Shen, Yang Sun, Joseph R. Cavallaro, Aida Vosoughi, and Yuanbin Guo
  IEEE Transactions on Circuits and Systems I - Regular Papers (TCAS-I), 2014. (Invited)
  Preprint version available at arXiv. http://arxiv.org/abs/1403.3759
  IEEEExplore version: download link
• Computer Vision Accelerators for Mobile Systems based on OpenCL GPGPU Co-Processing
  Guohui Wang, Yingen Xiong, Jay Yun, and Joseph R. Cavallaro
  Journal of Signal Processing Systems (JSPS), 2014. (Invited)
  Preprint version available at arXiv. http://arxiv.org/abs/1403.4238
  Final version (Springer website): download link
• Large-Scale MIMO Detection for 3GPP LTE: Algorithm and FPGA Implementation
  Michael Wu, Bei Yin, Guohui Wang, Chris Dick, Joseph R. Cavallaro, and Christoph Studer
  IEEE Journal of Selected Topics in Signal Processing (JSTSP), 2014.
  Preprint version available at arXiv. http://arxiv.org/abs/1403.5711
  IEEEExplore version: download link
• GPU Acceleration of a Configurable N-Way MIMO Detector for Wireless Systems
  Michael Wu, Bei Yin, Guohui Wang, Christoph Studer, and Joseph R. Cavallaro
  Journal of Signal Processing Systems (JSPS), 2014. (Invited)
  
• Implementation of a High Throughput 3GPP Turbo Decoder on GPU
  Michael Wu, Yang Sun, Guohui Wang, and Joseph R. Cavallaro
  Journal of Signal Processing Systems (JSPS), 2011.
  
• A Novel Design Of the High Speed Buffer and Video/audio Synchronization in High Resolution Digital Cinema System
  Guohui Wang, Zhenhua Zhu, Ke Zhang, Zhensong Wang
  High Technology Letters (In Chinese), Vol.9, 2008.
  
• Parallel VLSI Architecture for 3GPP LTE/LTE-Advanced Turbo Decoder
  Yang Sun, Guohui Wang, and Joseph R. Cavallaro
  submitted to IEEE Transactions on Signal Processing. (Under review)

[Conference Papers]
• [New!] OpenCL-Based Mobile GPGPU Benchmarking: Methods and Challenges
  Rotem Aviv, and Guohui Wang
  International Workshop on OpenCL (IWOCL), April, 2016.
  
• On the Performance of LDPC and Turbo Decoder Architectures with Unreliable Memories
  Joao Andrade, Aida Vosoughi, Guohui Wang, Georgios Karakonstantis, Andreas Burg, Gabriel Falcao, Vitor Silva, and Joseph R. Cavallaro
  48th IEEE Asilomar Conference on Signals, Systems, and Computers (ASILOMAR), November 2014.
  
• A High Performance GPU-based Software-defined Basestation
  Kaipeng Li, Michael Wu, Guohui Wang, and Joseph R. Cavallaro
  48th IEEE Asilomar Conference on Signals, Systems, and Computers (ASILOMAR), November 2014.
  
• Efficient Architecture Mapping of FFT/IFFT for Cognitive Radio Networks
  Guohui Wang, Bei Yin, Inkeun Cho, Joseph R. Cavallaro, Shuvra Bhattacharyy, and Jorma Takala
  IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), May 2014.
  
• A 3.8 Gb/s Large-scale MIMO Detector for 3GPP LTE-Advanced
  Bei Yin, Michael Wu, Guohui Wang, Chris Dick, Joseph R. Cavallaro, and Christoph Studer
  IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), May 2014.
  
• High Throughput Low Latency LDPC Decoding on GPU for SDR Systems
  Guohui Wang, Michael Wu, Bei Yin, and Joseph R. Cavallaro
  1st IEEE Global Conference on Signal and Information Processing (GlobalSIP), December 2013.
  Reference code
• Workload Analysis and Efficient OpenCL-based Implementation of SIFT Algorithm on a Smartphone
  Guohui Wang, Blaine Rister, and Joseph R. Cavallaro
  1st IEEE Global Conference on Signal and Information Processing (GlobalSIP), December 2013.
  
• HSPA+/LTE-A Turbo Decoder on GPU and Multicore CPU
  Michael Wu, Guohui Wang, Bei Yin, Christoph Studer, and Joseph R. Cavallaro
  47th IEEE Asilomar Conference on Signals, Systems, and Computers (ASILOMAR), November 2013.
  
• Highly Scalable On-the-Fly Interleaved Address Generation for UMTS/HSPA+ Parallel Turbo Decoder
  Aida Vosoughi, Guohui Wang, Hao Shen, Joseph R. Cavallaro, and Yuanbin Guo
  24th IEEE International Conference on Application-specific Systems, Architectures and Processors (ASAP), June 2013.
  
• Accelerating Computer Vision Algorithms Using OpenCL Framework on the Mobile GPU - A Case Study
  Guohui Wang, Yingen Xiong, Jay Yun, and Joseph R. Cavallaro
  IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), May 2013.
  
• A Fast and Efficient SIFT Detector using the Mobile GPU
  Blaine Rister, Guohui Wang, Michael Wu and Joseph R. Cavallaro
  IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), May 2013.
  
• Parallel Interleaver Architecture with New Scheduling Scheme for High Throughput Configurable Turbo Decoder (Finalist, best student paper award)
  Guohui Wang, Aida Vosoughi, Hao Shen, Joseph R. Cavallaro, and Yuanbin Guo
  IEEE International Symposium on Circuits and Systems (ISCAS), May 2013.
  
• Parallel Nonbinary LDPC Decoding on GPU
  Guohui Wang, Hao Shen, Bei Yin, Michael Wu, Yang Sun, and Joseph R. Cavallaro
  46th IEEE Asilomar Conference on Signals, Systems, and Computers (ASILOMAR), November 2012.
  
• Low Complexity Opportunistic Decoder for Network Coding
  Bei Yin, Michael Wu, Guohui Wang, and Joseph R. Cavallaro
  46th IEEE Asilomar Conference on Signals, Systems, and Computers (ASILOMAR), November 2012.
  
• GPGPU Accelerated Scalable Parallel Decoding of LDPC Codes
  Guohui Wang, Michael Wu, Yang Sun, and Joseph R. Cavallaro
  45th IEEE Asilomar Conference on Signals, Systems, and Computers (ASILOMAR), November 2011.
  
• High-throughput Contention-Free concurrent interleaver architecture for multi-standard turbo decoder
  Guohui Wang, Yang Sun, Joseph R. Cavallaro and Yuanbin Guo
  22nd IEEE International Conference on Application-specific Systems, Architectures and Processors (ASAP), September 2011.
  
• A Massively Parallel Implementation of QC-LDPC Decoder on GPU
  Guohui Wang, Michael Wu, Yang Sun, and Joseph R. Cavallaro
  9th IEEE Symposium on Application Specific Processor (SASP), June 2011.
  
• Multi-Layer Parallel Decoding Algorithm and VLSI Architecture for Quasi-Cyclic LDPC Codes
  Yang Sun, Guohui Wang, and Joseph R. Cavallaro
  IEEE International Symposium on Circuits and Systems (ISCAS), May 2011.
  
• FPGA Prototyping of A High Data Rate LTE Uplink Baseband Receiver
  Guohui Wang, Bei Yin, Kiarash Amiri, Yang Sun, Michael Wu, and Joseph R. Cavallaro
  43rd IEEE Asilomar Conference on Signals, Systems and Computers (ASILOMAR), November 2009.
  
  

[Posters]
• Parallel Interleaver Design for High Throughput Configurable Turbo Decoder
  2nd place winner best graduate student poster.
  Annual Rice University ECE Affiliates Day Conference, April 2013.
  
• Parallel Interleaver Design for High Throughput Configurable Turbo Decoder
  IEEE Texas Workshop on Integrated System Exploration (TexasWISE), March 2013.
  
• Low Energy Fast SIFT Detector on Heterogeneous Mobile Processors
  IEEE Texas Workshop on Integrated System Exploration (TexasWISE), March 2013.
  

[Patents]
• 6 US patent applications on computer vision and deep learning technologies.
  U.S. Patent Application. Filed by Snap Inc., 2017-2018.
  
• System and Method for a Turbo Decoder with Parallel Interleaver
  U.S. Patent Application. Filed by Huawei, October 2012.
  
• System and Method for Turbo Code Interleaved Address Generation
  U.S. Patent Application. Filed by Huawei, October 2012.
  
• System and Method for Contention-Free Memory Access
  U.S. Patent US8621160 B2. Filed by Huawei, December 2011. Granted, December 2013.
  (Also published as CN103262425A, WO2012079543A1)
  
• The Method, System and Device to Implement Video/audio Synchronization
  China Patent ZL200710120585.0. Filed in August 2007. Granted, September 2012.
• A Fast and High Performance Zooming Method for Multimedia Video
  China Patent ZL200710178188.9. Filed, 2007. Granted, October 2011.
  
• A copyright protection method and system for audio and video contents in digital cinema
  China Patent ZL200810114749.3. Filed in 2008. Granted, March 2010.
• A method of Watermark Generation and Detection for digital cinema Copyright Protection
  China Patent ZL200810103472.4. Filed in 2008. Granted, September 2010.

[Thesis]
• Ph.D thesis: "Design Space Exploration of Parallel Algorithms and Architectures for Wireless Communication and Mobile Computing Systems", Rice University, Houston, Texas, May, 2014.
• Master's thesis: "VLSI Architecture for High Definition Digital Cinema Playback System" (Abstract), Chinese Academy of Sciences, Beijing, China, June, 2008. Relative Project: Research and Implementation of DCI-Compliant 2K Digital Cinema Server (Jan.2006-Jan.2008, in ICT,CAS, Beijing, China).

Academic Acativities

• Teaching assitant:
  • ELEC 220 Fundamental of Computer Engineering (TA and lab instructor): Fall 2009, 2010, 2011, 2012.
  • ELEC 303 Random Signals: Fall 2009
  • ELEC 522 Advanced VLSI Design: Fall 2010
  • Technical notes: Catapult C Synthesis Work Flow Tutorial, (Tools: Catapult C, Xilinx ISE, System Generator, ModelSim)
• Paper reviewer:
  • Journal: JSSC, TSP, TCAS-I, TCAS-II, TPDS, TCSVT, TMM, TAES, TSC, CL, CAL, TECS, JSPS, CPE, JCST, JFCS, CSSP, SIVP, PARCO, JWCN, ADHOC, WINE, WIRE, JSCS, SRE, IJWI
  • Conference: ISCAS, ICASSP, ICC, GLOBECOM, GlobalSIP, EUSIPCO, GLSVLSI, ISITA, ASAP, SIPS, VTC, IWCMC, ISIEA

Useful Links

• Friends: Hui Wang, Bei Yin, Xiaozhu Lin, Yang Sun, Kia Amiri, Yang Zhao, Michael Wu

[Back to Top]
[Last Updated: 6/2018]