python main.py --hiddenSize 200 --epoch 1
/home/cc/flashnet/model_collection/5_block_prefetching/dataset/iotrace/
There are seven python files in this folders.
- cache.py: A cache simulator file, which includes a class about cache for cache initialization and execution.
- cachetest.py: Cache test for single-step and rolling prefetching, which includes data pre-processing and top-K encoding.
- data_process.py: Complete data processing file for both MSRC, HW and other datasets. It will save the trace as the stream form for model training and model test.
- main.py: Main function of our SGDP data prefetching framework. We can train and test the SGDP and its variants by python main.py based on default parameters.
- model.py: Our stream-graph neural network model and the subfunction for training, testing and prediction.
- pred_list_gene.py: Generate the local and global connection matrix and hybrid connection matrix. And executive the single-step and rolling data prefetching, output the HR and EPR for corresponding input trace.
- utils.py: Other functions for data processing and model building. Besides, there is a function for global stream-graph structure.