Papers
- CuPBoP: A framework to make CUDA portable
Ruobing Han, Jun Chen*, Bhanu Garg*, Jeffrey Young, Jaewoong Sim, Hyesoon KimPOSTER
PPOPP'23: 28th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming
- DynaaDCP: Dynamic Navigation of Autonomous Agents for Distributed Capture Processing
Sam Jijina, Ramyad Hadidi, Jun Chen, Zhen Jiang, Ashutosh Dhekne, Hyesoon Kim
DOSSA-4 @ HPCA'22
Paper
Projects
CuPBoP-AMD is a CUDA translator that translates CUDA programs at NVVM IR level to HIP-compatible IR that can run on AMD GPUs.
CuPBoP is proposed to execute CUDA on non-NVIDIA devices without relying on any portable programming languages. Compared with existing work that executes CUDA on non-NVIDIA devices, CuPBoP does not require manual modification of the CUDA source code. I worked on CUDA runtime for x86 architecture and used LLVM to manipulate bitcode generated from frontend compilation. I helped with benchmarking and debugging backend compiler issues and the runtime system.
Created a course and professor search for students interested in the grade point average or letter grade percentages of offered courses
Implemented OpenDroneMap with gRPC for distributed 3D reconstruction for the DynaaDCP: Dynamic Navigation of Autonomous Agents for Distributed Capture Processing
Benchmark TimescaleDB with Yahoo Finance historical data
A Flask application that collects the data to determine the number of second genration athletes per a collegitate sports team