Skip to content

Code release of "DeepOHeat: Operator Learning-based Ultra-fast Thermal Simulation in 3D-IC Design", DAC 2023. https://arxiv.org/pdf/2302.12949

License

Notifications You must be signed in to change notification settings

Cadence-Celsius/DeepOHeat

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepOHeat (DAC 2023)

DeepOHeat: Operator Learning-based Ultra-fast Thermal Simulation in 3D-IC Design [Paper]

Thermal-aware floorplanning is particularly important in designing 3D integrated circuits. To reduce turnaround time in design optimization loops and avoid collecting data in substantial costs, we propose DeepOHeat with the following features:

  • For the first time, DeepOHeat solves a family of heat equations under various 3D-IC configurations in an end-to-end, data-free scheme.
  • We encode 3D-IC configurations using a modular approach: stacked cuboidal geometry, boundary conditions, 2d/3d power maps and full chip material conductivity.
  • We achieve satisfactory accuracy when benchmarking DeepOHeat against Cadence Celsius Thermal Solver.



Figure 1. DeepOHeat framework.

Getting Started

Python 3.8.16 was used to develop this repo, to install dependencies

pip install -r requirements.txt

To reproduce examples in our paper or directly use pretrained checkpoints, please refer to 2d_power_map and multi_htc_bc.

  • In particular, prototype.py directly runs inference on the 10 test cases we showcased in the paper, at Section V.A.
  • Side by side comparison against Cadence Celsius is in the benchmark folder of each experiment's logging directory, if provided.

The single_case_pinns directory contains examples solving single instances of various 3D-IC configurations.

  • complex_geometry showcases a geometry of two stacked cuboids with a volumetric power defined in the middle of the bottom cuboid.
  • quarter_power and uniform_power showcase 2d surface power on top with different sizes.
  • volumetric_power showcases a single cuboidal geometry with a volumetric power.

Showcase



Figure 2. Benchmark DeepOHeat against Cadence Celsius. Top -> bottom: Celsius, DeepOHeat, point-wise error.

Citation

If you find this repo useful, please cite our paper.

@inproceedings{liu2023deepoheat,
  title={DeepOHeat: operator learning-based ultra-fast thermal simulation in 3D-IC design},
  author={Liu, Ziyue and Li, Yixing and Hu, Jing and Yu, Xinling and Shiau, Shinyu and Ai, Xin and Zeng, Zhiyu and Zhang, Zheng},
  booktitle={2023 60th ACM/IEEE Design Automation Conference (DAC)},
  pages={1--6},
  year={2023},
  organization={IEEE}
}

License

DeepOHeat is provided under MIT License, please see LICENSE for full license text.

Contact

If you have any question regarding the code or the paper, please contact ziyueliu@ucsb.edu.

Acknowledgement

This is a joint work between Zhang's group at UCSB and Celsius R&D team at Cadence Design Systems.

About

Code release of "DeepOHeat: Operator Learning-based Ultra-fast Thermal Simulation in 3D-IC Design", DAC 2023. https://arxiv.org/pdf/2302.12949

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages