This converter tells you how to convert a text edge tuple into binary CSR format in parallel

• Create a folder to store dataset: mkdir -p sample_dataset; cd sample_dataset
• Download a text tuple list: wget https://snap.stanford.edu/data/bigdata/communities/com-orkut.ungraph.txt.gz
• Uncompress the tuple list file: gzip -d com-orkut.ungraph.txt.gz
• To improve processing parallelism, you should split the file into a number of smaller files with our provided bash script: ../split_rename.bash com-orkut.ungraph.txt
• Count number of smaller files we get from prior step: ls -l ./com-orkut.ungraph.txt-split* | wc -l
• Convert those files into binary files: ../tuple_to_bin.multithread/text_to_bin.bin ./com-orkut.ungraph.txt-split 201 32 1. Here you can type ../tuple_to_bin.multithread/text_to_bin.bin to check the meanings of all input parameters.
• Convert those edge tuples into binary CSR: ../multi_bin_to_2d_csr/multi_bin_to_2d_csr.bin ./com-orkut.ungraph.txt-split 201 1 1 32. Again, you can type ../multi_bin_to_2d_csr/multi_bin_to_2d_csr.bin to see the meaning of each input parameter.