Download

Download the Dataset

Based on your target tasks, download Routability Features (for congestion and DRC) or IR Drop Features (for IR drop) or Graph Features (for congestion and net delay).

Google Drive

Baidu Netdisk

Decompress and Preprocess

• Routability/IR Drop Features:

  (1) Decompress with scripts in the script dir

  python decompress_routability.py

  or

  python decompress_IR_drop.py

  Make sure your directory structure is the same as the one in Google Drive or Baidu Netdisk and you are using the latest version of the script in the drive.

  This may takes at least 15 minutes and 200G storage space, please be patient.

  (2) Run preprocessing script to generate training set for corresponding tasks. Specify your task with option: congestion/DRC/IR_drop.

  python generate_training_set.py --task [congestion/DRC/IR_drop] --data_path [path_to_decompressed_dataset] --save_path [path_to_save_output]

  (3) Now, you can set up your own model for training or use the tutorial code from our tutorial page and github repository https://github.com/circuitnet/CircuitNet.

• Graph Features:

  (1) Decompress with tar: tar -xf PATH_TO_THE_FILE

  (2) Construct graph with our sample code. (coming soon)

Raw Data

We provide netlist and LEF/DEF for user-defined feature extraction. The sample code for feature extraction has been released.