CARVIEW

MOTORHOMES

Select Language

HTTP/2 200 server: GitHub.com content-type: text/html; charset=utf-8 last-modified: Sun, 15 Jun 2025 04:03:01 GMT access-control-allow-origin: * etag: W/"684e45f5-2dfa" expires: Tue, 30 Dec 2025 14:27:05 GMT cache-control: max-age=600 content-encoding: gzip x-proxy-cache: MISS x-github-request-id: 5BEA:2680BD:A25854:B6527D:6953DEE0 accept-ranges: bytes age: 0 date: Tue, 30 Dec 2025 14:17:05 GMT via: 1.1 varnish x-served-by: cache-bom-vanm7210048-BOM x-cache: MISS x-cache-hits: 0 x-timer: S1767104225.376866,VS0,VE219 vary: Accept-Encoding x-fastly-request-id: 8cc804e54e92f62adbb8b165d885031a691d3bfd content-length: 4345 Siyuan Zhang | the Ohio State University

Siyuan Zhang

PhD Student
the Ohio State University
zhang.15333@osu.edu, ottozsy@outlook.com

About Me

I am a PhD student @ the Ohio State University, proudly advised by Prof DK Panda and Prof Hari Subramoni @ NOWLAB.

Research Interests

HPC : MPI, GPU, HPC4AI
MLSys

Experience

Research Assistant @ NOWLAB

TBD

Research Assistant @ L.I.O.N. & NeuDim

Co-author NEPHELE: A Neural Platform for Highly Realistic Cloud Radiance Rendering, as the second author.
- Reduce hash grid size by 8.7 times by introducing Perfect Spatial Hashing into NGP.
- Develop a multi-GPU, multi-node extrem low latency NGP rendering framework, achieving 50 FPS @ 1080 resolution for 12 NGP on 4 GPUs in one node and 37 FPS @ 1080 resolution for 12 NGP on 4 GPUs in 2 nodes using CUDA awared MPI.
Research in workloads of DNN Training Jobs in HPC

Student Cluster Competition (SCC) @ Geekpie_HPC Shanghaitech University

High Performance Computing (HPC) related research. Continually preparing and aiming for the top-tier student supercomputing competitions, including ASC, ISC and SC. Mainly in charge of HPC tasks related with AI, MPI and CUDA.

ASC22 @ USTC :
- Manage hardwares under restricted power constraints (3KW, 3 nodes, 8 GPUS).
- Use DeepSpeed to accelerate of YUAN A Large Language Model Pretrain Model.
- Fine-tuned HPC benchmarks HPCG and HPL.
ISC22 @ Online :
Profile and optimize ICON a MPI Program for earth system model (3rd place).
SC21 @ Online (2nd place):
- Reproduce the correctness of a SC research ramBLe (1st place).
- Run and tuning a multi-node multi-GPU DNN image segmentation application CosmicTagger. (2nd place).
ISC21 @ Online:
- Modify a GPU optimized HPCC benchmark
- Research on MPI and OpenMP balance for WRF
- Contributed for collective_profiler a set of tools and libraries to profile Alltoallv MPI calls.

Teaching Assistant

CS171: Computer Graphics ShanghaiTech University

Skills

Programming Languages

C/C++, CUDA, Python

Frameworks, Libs

OpenMP, MPI, Pytorch, OpenGL, OpenCV, NCCL

Tools

Nsight System, Nsight Compute, Intel Vtune, Intel Trace Analyser and Collector, IPM.

Publications

TPDS

Critique of “A parallel framework for constraint-based Bayesian network learning via Markov blanket discovery” by SCC team from ShanghaiTech University

Guancheng Li, Songhui Cao, Chuyi Zhao, Siyuan Zhang, Yuchen Ji, Haotian Jing, Zecheng Li, Jiajun Cheng, Yiwei Yang, Shu Yin* (*Corresponding authors)

IEEE Transactions on Parallel and Distributed Systems

PDF BibTex

NEPHELE: A Neural Platform for Highly Realistic Cloud Radiance Rendering

Haimin Luo, Siyuan Zhang, Fuqiang Zhao, Haotian Jing, Penghao Wang, Zhenxiao Yu, Dongxue Yan, Junran Ding, Boyuan Zhang, Qiang Hu, Shu Yin, Lan Xu, JIngyi Yu

PDF BibTex proceeding

Original Source | Taken Source