Research Note
High Performance Computing
Proficiency
Proficient
Description
- Large Eddy Simulation (LES) execution and debugging (DAM model), ~46k core-hours per run
- Code extension: adding diagnostic output variables, new forcing logic, I/O cadence optimization
- Compute resource planning: designing tractable sensitivity plans within compute budgets
- Idealized aquaplanet simulations
Used In
- — DAM model 3D LES, 46k core-hours/run
- Undergraduate Research @ Peking University & Caltech/UCR — idealized aquaplanet simulations
- Walker Circulation Dynamics @ PKU — Tianhe-II 48-PE + CAM3 aqua-planet
- DSCOVR Inverse Problem + Regularization Methods @ Caltech / UCR — (remote workstation) bootstrap + 5-language
- — Lawrencium production runs
Research-grade signature capabilities
PKU undergraduate HPC infra mastery (2020-2021)
- Tianhe-II NSCC-GZ: 3 accounts ((undergraduate Tianhe-II account) primary + 2 more Walker variants);
yhrun -p work -n 4848-PE SLURM (rare for an undergraduate) - CAM3 aqua-planet 30-yr spin-up runs × 10 cases (D60-180 + Wonly + Conly + Aqua, each case a 30-year integration)
- 27,882 MATLAB+NCL analysis lines + 5-stage HPC pipeline architecture (CAM3 aqua-planet → Tianhe-II runs → 25 cal_X MATLAB diagnostics → NCL stream function + ω analysis → Python figure polish)
UC Berkeley PhD HPC mastery (2022-2025)
- Lawrencium LBNL cluster: 8 production runs × ~46k core-hours each ≈ 44 days HPC wall time engineering plan
- DAM quasi-2D arrays Lx 20km-50000km scan (Phase 4.5 killer-figure convergence)
- 115,363 Jupyter notebook lines / 62 ipynb / 3,793 code cells aggregate (PhD analysis code)
Cross-language HPC tooling
- Fortran 77/90: source-code-level understanding of NCAR CAM3 v3.1 source tree (615 .F90 files) + DAM LES source (40+ modules)
- Python/MATLAB/R/NCL/IDL/Mathematica: cross-language scientific computing (Caltech DSCOVR 5-language fluency)
- remote workstation CUDA bootstrap (Caltech, Windows CUDA incompatibility → manual CUDA driver + Anaconda bootstrap)