DOAJ Open Access 2025

PyPOD-GP: Using PyTorch for accelerated chip-level thermal simulation of the GPU

Neil He Ming-Cheng Cheng Yu Liu

Abstrak

The rising demand for high-performance computing (HPC) has made full-chip dynamic thermal simulation in many-core GPUs critical for optimizing performance and extending device lifespans. Proper orthogonal decomposition (POD) with Galerkin projection (GP) has shown to offer high accuracy and massive runtime improvements over direct numerical simulation (DNS). However, previous implementations of POD-GP use MPI-based libraries like PETSc and FEniCS and face significant runtime bottlenecks. We propose a PyTorch-based POD-GP library (PyPOD-GP), a GPU-optimized library for chip-level thermal simulation. PyPOD-GP achieves over 23.4× speedup in training and over 10× speedup in inference on a GPU with over 13,000 cores, with just 1.2% error over the device layer.

Topik & Kata Kunci

Computer software

Penulis (3)

Neil He

Ming-Cheng Cheng

Yu Liu

Format Sitasi

APA MLA BibTeX

He, N., Cheng, M., Liu, Y. (2025). PyPOD-GP: Using PyTorch for accelerated chip-level thermal simulation of the GPU. https://doi.org/10.1016/j.softx.2025.102147

Akses Cepat

PDF tidak tersedia langsung

Cek di sumber asli →

Lihat di Sumber doi.org/10.1016/j.softx.2025.102147

Informasi Jurnal

Tahun Terbit: 2025
Sumber Database: DOAJ
DOI: 10.1016/j.softx.2025.102147
Akses: Open Access ✓