Skip to content

Latest commit

 

History

History
52 lines (40 loc) · 2.07 KB

README.md

File metadata and controls

52 lines (40 loc) · 2.07 KB
Raw

Figure 5,6,7 : Experimental Scalibility Results for Resilience and Responsibility Problems

Instructions to Reproduce Figures 5,6,7

  1. Make sure all prerequisites are satisfied as specified in README.md

  2. Download TPC-H data

    1. Clone the dbgen repository
        cd data/tpch
    
    git clone https://github.com/electrum/tpch-dbgen.git

    cd ../..
    1. Build dbgen by running
        cd data/tpch/tpch-dbgen
    make
    1. Run the tpch-data-gen.sh script to generate the TPC-H data at various scale factors
        bash SIGMOD_Reproducibility/Fig-5,6,7/tpch-data-gen.sh
    1. You should see 13 folders in data/tpch/data, corresponding to data generated for 13 scale factors.

  3. Test Gurobi install by running gurobi_cl

  4. Run the experiments

        bash SIGMOD_Reproducibility/Fig-5,6,7/run_experiments.sh

4(b). Optionally just run the experiments for Fig 5 and 7 for smaller sizes (takes <3 hours to terminate) bash SIGMOD_Reproducibility/Fig-5,6,7/run_experiments_small.sh

4(c). Optionally just run the experiments for Fig 6 on tpch data (takes <30min to terminate) bash SIGMOD_Reproducibility/Fig-5,6,7/run_experiments_tpch.sh

  1. Visualize the output: We have a Jupyter notebook sigmod_expts.ipynb code to read the experimental output and generate the plots by calling methods from ../../src/resilience_responsibility_solver/plot/.

  2. (Optional) The code generates some intermeadiate files, like pickle files storing various data instances that are tested, log files, linear program files. To delete this, you can run

        bash SIGMOD_Reproducibility/Fig-5,6,7/clean_intermediate_data.sh

The default notebook uses the reference experimental data to start with, but this can be toggled by setting use_reference_data = False. To use the data generated in smaller sized experiments set use_small_expt = True.