DuST: Fast Dynamic IR-Drop Prediction with Dual-path Spatial-Temporal Attention

This is the implementation of the DATE'25 paper: Fast Dynamic IR-Drop Prediction with Dual-path Spatial-Temporal Attention

The model architecture is as shown in the following figure.

Requirements

Dependencies are listed in requirements.txt and can be installed by:

pip install -r requirements.txt

Dataset

Please refer to CircuitNet to download and extract the dataset for dynamic ir drop prediction task. Put the dataset in the train_data/ folder. The directory structure should be like this:

├── train_data
│   ├── feature
│   │   ├── <feature_map_1>.npy
│   │   ├── <feature_map_2>.npy
│   │   ├── ...
│   ├── label
│   │   ├── <label_map_1>.npy
│   │   ├── <label_map_2>.npy
│   │   ├── ...

The dataset is indexed by the files under index/ folder.

Configuration

Please refer to utils/configs.py to modify the configurations, where the default parameters are used in our experiments.

Usage

The pretrained model can be downloaded here. We provide a pretrained model in pretrained/model_iters_176950.pth and can be downloaded by:

cd pretrained
bash download.sh

Please refer to train.py to perform the training. The testing is included in test.py.

# for Training
python train.py --model_type IRNetDual

# for Testing
python test.py --model_type IRNetDual --pretrained pretrained/model_iters_176950.pth --result_dir results/test

Citation

If you find our work useful in your research, please consider to cite:

@inproceedings{fu2025ir_predict,
    author={Fu, Bangqi and Liu, Lixin and Wang, Qijing and Wang, Yutao and Wong, Martin D. F. and Young, Evangeline F. Y.},
    booktitle={Proceedings of the 2025 IEEE/ACM Design, Automation and Test in Europe Conference},
    title={Fast Dynamic IR-Drop Prediction with Dual-path Spatial-Temporal Attention},
    year={2025},
}