Luanzheng "Lenny" Guo

I am joining the Pacific Northwest National Laboratory HPC Group as a postdoc researcher, working in the research area between HPC and Machine Learning. Previously, I obtained my Ph.D. degree in Electrical Engineering & Computer Science from the University of California-Merced. My Ph.D. research focuses on system resilience and reliability in large-scale parallel HPC systems. My research interests lie in the broad intersection of HPC system, Cybersecurity, and Data Science.

nEWS

8/2020
A paper is accepted by IISWC'20!
7/2020
Lenny is serving as a Technical Program Session Chair for PEARC'20!
7/2020
Lenny successfully defended his PhD!
2/2020
Lenny is selected as an NSF Trusted CI fellow for the 2020 class by the NSF Trusted Cyberinfrastructure program!
2/2020
A paper is accepted by ISC'20!

Achievements

2020
NSF Trusted CI Fellow for the 2020 class
2019
UC Merced Graduate Dean's Dissertation Fellowship
2019
UC Merced GSA Travel Award for SC’19
2019
IEEE CS TCHPC Travel Award for SC’19
2019
UC Merced Graduate Fellowship Incentive Program Award
2019
IEEE TCPP Travel Award for IPDPS’19
2018
Research highlighted by HPCwire in its “What’s new in HPC research?”
2018
SC’18 Outstanding Student Volunteer
2018
SC’18 Outstanding Lightning Talk
2018
Outstanding Reviewer Award by Elsevier [Certificate]
2017
ACM SIGHPC Travel Award to attend the 50th celebration of the Turing Award
2016
Best poster finalist in SC’16
2015 - 2017
Bobcat Fellowship at UC Merced

Publications

MATCH: An MPI Fault Tolerance Benchmark Suite

Luanzheng Guo, Giorgis Georgakoudis, Konstantinos Parasyris, Ignacio Laguna, and Dong Li.

in IEEE International Symposium on Workload Characterization (IISWC'20), 08/2020.

Pdf Bibtex

@inproceedings { LGuo : iiswc_20, author = "Luanzheng Guo, Giorgis Georgakoudis, Konstantinos Parasyris, Ignacio Laguna, and Dong Li", title = "{MATCH: An MPI Fault Tolerance Benchmark Suite}", booktitle = "{IEEE International Symposium on Workload Characterization (IISWC'20)}", month = "Aug", year = "2020", }

Code

Reinit++: Evaluating the Performance of Global-Restart Recovery Methods For MPI Fault Tolerance

Giorgis Georgakoudis, Luanzheng Guo and Ignacio Laguna.

in ISC HPC Conference (ISC'20), 02/2020.

Pdf Bibtex

@inproceedings { LGuo : phdshowcase_19, author = "Giorgis Georgakoudis, Luanzheng Guo and Ignacio Laguna", title = "{Reinit++: Evaluating the Performance of Global-Restart Recovery Methods For MPI Fault Tolerance}", booktitle = "{ISC HPC Conference (ISC'20)}", month = "Feb", year = "2020", }

Code

PhD-Showcase: Characterization and Modeling of Error Resilience in HPC Applications

Luanzheng Guo and Dong Li.

in ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC'19), 11/2019.

Pdf Bibtex

@inproceedings { LGuo : phdshowcase_19, author = "Luanzheng Guo and Dong Li", title = "{Characterization and Modeling of Error Resilience in HPC Applications}", booktitle = "{ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC’19)}", month = "Nov", year = "2019", }

Code

PhD-Showcase: Characterization and Modeling of Error Resilience in HPC Applications

Luanzheng Guo and Dong Li.

in IEEE International Parallel and Distributed Processing Symposium (IPDPS'19), 05/2019.

Pdf Bibtex

@inproceedings { LGuo : phdshowcase_19, author = "Luanzheng Guo and Dong Li", title = "{Characterization and Modeling of Error Resilience in HPC Applications}", booktitle = "{IEEE International Parallel and Distributed Processing Symposium (IPDPS'19)}", month = "May", year = "2019", }

Code

Paper: MOARD: Modeling Application Resilience to Transient Faults on Data Objects

Luanzheng Guo and Dong Li.

in IEEE International Parallel and Distributed Processing Symposium (IPDPS'19), 05/2019.

Pdf Bibtex

@inproceedings { LGuo : moard_19, author = "Luanzheng Guo and Dong Li", title = "{MOARD: Modeling Application Resilience to Transient Faults on Data Objects}", booktitle = "{IEEE International Parallel and Distributed Processing Symposium (IPDPS'19)}", month = "May", year = "2019", }

Code

Paper: FlipTracker: Understanding Natural Error Resilience in HPC Applications Research highlighted by HPCwire

Luanzheng Guo, Dong Li, Ignacio Laguna, and Martin Schulz.

in ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC'18), 11/2018.

Pdf Bibtex

@inproceedings { LGuo : fliptracker_18, author = "Luanzheng Guo and Dong Li and Ignacio Laguna and Martin Schulz", title = "{FlipTracker: Understanding Natural Error Resilience in HPC Applications}", booktitle = "{ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC’18)}", month = "Nov", year = "2018", }

Code

Paper: Indoor Frame Recovering Via Line Segments Refinement and Voting.

Jun Chu, Lenny GuoLu, Lingfeng Wang, Chunhong Pan, and Shiming Xiang.

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2013), 26/05/2013.

Pdf Bibtex

@inproceedings { JChu : frame_recovery_13, author = "Jun Chu and Lenny GuoLu and Lu Wang", title = "{Indoor Frame Recovering Via Line Segments Refinement and Voting}", booktitle = "{IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP’2013)}", month = "May", year = "2013", }

Code-link1 Code-link2

Paper: Chessboard corner detection under Image physical coordinate.

Jun Chu, Lenny GuoLu, and Lu Wang.

Optics and Laser Technology, 48(0): 599-605, 2013.

Pdf Bibtex

@article { JChu : physical_coordinate_13, author = "Jun Chu and Lenny GuoLu and Lu Wang", title = "{Chessboard corner detection under Image physical coordinate}", journal = "Optics and Laser Technology", volume = 48, number = 0, pages = "599-605", year = {2013}, }

Code-link1 Code-link2

Posters

Understanding Use of ULFM in MPI Fault Tolerance

Luanzheng Guo and Giorgis Georgakoudis and Ignacio Laguna.

in Lawrence Livermore National Laboratory Summer Poster Symposium 2019.

Pdf Bibtex

@inproceedings { LGuo : llnl_poster_19, author = "Luanzheng Guo and Giorgis Georgakoudis and Ignacio Laguna", title = "{Understanding Use of ULFM in MPI Fault Tolerance}", booktitle = "{Lawrence Livermore National Laboratory Summer Poster Symposium}", month = "Aug", year = "2019", }

Code

Which Fault Injection Tool Should We Use for GPU Programs?

Luanzheng Guo and Ignacio Laguna.

in Lawrence Livermore National Laboratory Summer Poster Symposium 2018.

Pdf Bibtex

@inproceedings { LGuo : llnl_poster_18, author = "Luanzheng Guo and Ignacio Laguna", title = "{Which Fault Injection Tool Should We Use for GPU Programs?}", booktitle = "{Lawrence Livermore National Laboratory Summer Poster Symposium}", month = "Aug", year = "2018", }

Code

Understanding the Resilience of Fundamental Data Types

Luanzheng Guo and Ignacio Laguna.

in Lawrence Livermore National Laboratory Summer Poster Symposium 2017.

Pdf Bibtex

@inproceedings { LGuo : llnl_poster_17, author = "Luanzheng Guo and Ignacio Laguna", title = "{Understanding the Resilience of Fundamental Data Types}", booktitle = "{Lawrence Livermore National Laboratory Summer Poster Symposium}", month = "Aug", year = "2017", }

Code

Understanding Fault Tolerance Patterns of Algorithms via Application-Level Fault Injection

Luanzheng Guo, Ignacio Laguna, and Martin Schulz.

in Lawrence Livermore National Laboratory Summer Poster Symposium 2016.

Pdf Bibtex

@inproceedings { LGuo : llnl_poster_16, author = "Luanzheng Guo and Ignacio Laguna and Martin Schulz", title = "{Understanding Fault Tolerance Patterns of Algorithms via Application-Level Fault Injection}", booktitle = "{Lawrence Livermore National Laboratory Summer Poster Symposium}", month = "Aug", year = "2016", }

Code

Understanding Ineffectiveness of Application-Level Fault Injection. Best Poster Finalist

Luanzheng Guo, Jing Liang, and Dong Li.

in ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC'16).

Pdf Bibtex

@inproceedings { LGuo : fault_injection_16, author = "Luanzheng Guo and Jing Liang and Dong Li", title = "{Understanding Ineffectiveness of Application-Level Fault Injection}", booktitle = "{ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (SC’16)}", month = "Nov", year = "2016", }

Code

Projects

“Building the Detailed Dynamic Dependency Graph (DDDG) for Applications” Advisor: Prof. Dong Li, Ignacio Laguna.

view details

Constructing the DDDG graph for an entire application instead of only one function, one module, or one loop, in which the values for registers and memories, and the dynamic control flows are visible.

Pdf

“Identification of the error resilience logic for machine learning like algorithms on large scale system”, Mentor: Ignacio Laguna, Prof. Dong Li.

view details

The target is to recognize the inner logic of some algorithms that why they can tolerate underneath faults of large scale system derived from device worn, cosmos radiation like neutrons, and temperature. And then coming up with a methodology to automatically detect the error resilience logic of the algorithms.

Pdf

“Analysis on application-level resilience on large scale system”, advisor: Prof. Dong Li.

view details

soft errors (transient faults) due to reasons like cosmic rays are already making an impact in industry, and so many internet companies suffer from it. Generally, the damage brought by soft errors can be rolled back with great overhead, so what should we do to solve the problem? One feasible way is to protect the data and make the program more robust. Our goal is to propose a metric to evaluate the vulnerability of data within specific program, which will be used as the criterion to estimate the degree of protection on each data.

Pdf

“Target Tracking Based on Context-aware and Sparse Representation”, advisor: Prof. Lingfeng Wang.

view details

prior research on classifying human beings on depth information.

1) read papers related to the specific research area and make a conclusion on existing methods; 2) try to propose a novel model to solve the problem from a new prospect.

Pdf

“Gesture Recognition based on 2D and 3D Information” (course project)

view details

Identifying different gestures based on their 2D and 3D information.

1) fetching the 2D and 3D information of gestures with Kinect; 2) extracting the palm area out of the background based on depth discrimination; 3) smoothing the 3D mesh and extract the outline of gestures; 4) calculating the convex-hull for gestures based on their outline; 5) using an Adaboost classifier to recognize different gestures.

Pdf

“Urban building modeling by integrating 2D and 3D information”, advisor: Prof. Chunhong Pan.

view details

Reasoning the framework of Indoor Scenes from a single image.

1) conducted some researches into the most essential part of indoor scene recognition, known as indoor frame recovery; 2) address difficulties such as illumination variations, weak boundaries and partial occlusions; 3) proposed an approach based on line segment refinement and voting; 4) refined detected line segments by revising, connecting, and adding operations; 5) utilized an iterative voting mechanism to select from refined line segments, where a cross ratio constraint is enforced to build the frame.

Pdf

“Theories of Linear Reconstruction Natural Scenes Based on Spatial Invariants”, advisor: Prof. Jun Chu.

view details

calibrating the panorama cameras stably under complex scenes. (Coding with Matlab)

1) considered approaches based on chessboard panel; 2) proposed two algorithms of chessboard corner detectors based on a round template and a circular template, respectively, under image physical coordinate; 3) the physical coordinates allowed detected chessboard corners to reach the sub-pixel level in only one step; 4) experiment results prove that our algorithms can achieve better results in complex scenes.

Pdf

“Key techniques for modeling the lunar surface for exploration and patrol”, advisor: Prof. Ouyang Ziyuan.

view details

finding a stable and efficient approach to calibrate the cameras fixed on the mobile robot. (Coding with C++)

1) there are four cameras, two panorama cameras and two navigation cameras. For the navigation cameras, I fixed their focal lengths. But for the panorama camera, the focal length could not be fixed; 2) discovered a problem that the calibration parameters of the panorama cameras could vary, as the robot is moving; 3) to address that, each time before taking pictures in 360 degrees, I stopped the robot and calibrated the panorama cameras with a fixed chessboard panel on the robot.

Pdf

Education

PhD Student in Computer Science, University of California, Merced, U.S., Current GPA: 3.8/4.0
Aug. 2015-present
M.S. in Vision and Pattern Recognition, Nanchang Hangkong University, China, GPA: 3.5/4.0
Sep. 2011- Jun. 2014

B.S. in Computer Science, Nanchang Hangkong University, China, GPA: 3.2/4.0
Sep. 2007- Jun. 2011

Activities

Student Volunteer Lead at SC’19, Denvo, CO (Winter, 2019)

Student Volunteer Lead at SC’18, Dallas, TX (Winter, 2018)

Student Volunteer at SC’17, Denvo, CO (Winter, 2017)

Traveling Scholarship to join the ACM 50th Celebration of the Turing Award

Student Volunteer at SC’16, Salt Lake City, U.S (Fall, 2016)

Attendee at SC’15, Austin, TX, U.S (Fall, 2015)

References

Dong Li, Associate Professor, University of California, Merced.

Email: dli35[at]ucmerced.edu
Ignacio Laguna, Computer Scientist, LLNL.

Email: lagunaperalt1[at]llnl.gov
Martin Schulz, Chief Professor, Technical University of Munich.

Email: schulzm[at]in.tum.de
Jun Chu, Professor, Nanchang Hangkong Unviersity.

Email: chujun[at]nchu.edu.cn
Chunhong Pan, Professor, Chinese Academy of Sciences.

Email: chpan[at]nlpr.ia.ac.cn
Shiming Xiang, Professor, Chinese Academy of Sciences.

Email: smxiang[at]nlpr.ia.ac.cn