|
My research interests lie in the general area of high-performance computing, particularly in parallel programming and algorithms, with the focus on graph algorithms and applied numerical linear algebra. I am an EECS PhD student at UC Berkeley, co-advisded by Jim Demmel and Aydın Buluç. I am associated with the BeBOP and the PASSION groups. I received my Bachelor's degree in Computer Engineering from UIUC in 2022, with Highest Honors and with a minor in Mathematics. I have had the privilege to work with Prof. Wen-mei Hwu, Prof. Rakesh Nagi, and Prof. Jinjun Xiong at Illinois. If you would like to learn more about me, please see my résumé or contact me via email. |
|
|
|
[3] Yen-Hsiang Chang, Aydın Buluç, James Demmel Preprint on arXiv arxiv | code The approximate minimum degree algorithm is widely used before numerical factorization to reduce fill-in for sparse matrices. While considerable attention has been given to the numerical factorization process, less focus has been placed on parallelizing the approximate minimum degree algorithm itself. In this paper, we explore different parallelization strategies, and introduce a novel parallel framework that leverages multiple elimination on distance-2 independent sets. Experimental results show that we achieve up to an 8.30x speedup using 64 threads over the state-of-the-art sequential implementation in SuiteSparse. |
|
[2] Mohammad Almasri*, Yen-Hsiang Chang*, Izzat El Hajj, Rakesh Nagi, Jinjun Xiong, Wen-mei Hwu (*Equal contribution) 32nd International Conference on Parallel Architectures and Compilation Techniques (PACT'23) paper | arxiv | code Instead of the breadth-first approaches used in prior GPU works, we propose to parallelize maximal clique enumeration on GPUs by performing depth-first traversal using the Bron-Kerbosch algorithm. We also propose a worker list for dynamic load balancing, as well as partial induced subgraphs and a compact representation of excluded vertex sets to regulate memory consumption. Our evaluation shows that our implementation on a single GPU outperforms the state-of-the-art parallel CPU implementation by a geometric mean of 4.9X (up to 16.7X), and scales efficiently to multiple GPUs. |
|
[1] Yen-Hsiang Chang, Jianhao Pu, Jinjun Xiong, Wen-mei Hwu 2021 BenchCouncil International Symposium on Benchmarking, Measuring and Optimizing (Bench'21) paper | arxiv | code | presentation We propose MLHarness, a scalable benchmarking harness system for MLCommons Inference developed on top of the MLModelScope system with three distinctive features: (1) it codifies the standard benchmark process as defined by MLCommons Inference including the models, datasets, DL frameworks, and software and hardware systems; (2) it provides an easy and declarative approach for model developers to contribute their models and datasets to MLCommons Inference; and (3) it includes the support of a wide range of models with varying inputs/outputs modalities so that we can scalably benchmark these models across different datasets, frameworks, and hardware systems. |
|
|
This is an academic website for Yen-Hsiang Chang to share his experiences, projects, publications. |