Skip to content

Latest commit

 

History

History
26 lines (18 loc) · 1.41 KB

README.md

File metadata and controls

26 lines (18 loc) · 1.41 KB
Raw

CODAG

Source code for the paper CODAG

Dataset

The datasets used in the paper CODAG can be downloaded from this link (https://drive.google.com/drive/folders/1q_3hCQz2dBGQw7VowBNANcVJ-6qvmDha?usp=sharing). Make sure both <dataset_name>_comp.bin and <dataset_name>_blk_offset.bin are downloaded for Deflate

Instruction

First, CODAG decompressors for RLE v1, RLE v2, and Deflate can be simply generated by calling make clean && make

Once all decompressors are generated, you can use the following commend to run RLE v1 or RLE v2 decompression. ./build/exe/src/<ENCODING_TECHNIQUE>/<ENCODING_TECHNIQUE>.cu.exe -f <ORIG_DATASET> -t <DATATYPE BYTE SIZE> .

For example, to run RLE v1 decompression on MC0 dataset with uint64_t datatype, the commend would be. ./build/exe/src/rlev1/rlev1.cu.exe -f col0.bin -t 8

To run Deflate decompression, the datasets are firsted compressed by deflate_compressor.py or simply use the compressed datasets from the link for the dataset.

Once, compressed datasets are ready, use the following commend to call CODAG Deflate decompression. ./build/exe/src/zlib/zlib.cu.exe <DATASET_NAME>_comp.bin (decompressed file) <DATASET_NAME>_blk_offset.bin <CHYNK_SIZE>. For the provided compressed dataset, CHUNK_SIZE should be 131072 (128KB).