Huffman-Compression

Huffman encoding and decoding with simple Serial and Parallel implementations for CPU and GPU (cuda)

Algorithm

1. Compression

* Read the input data
* Get the frequencies for each symbol
* Build a tree from the frequencies
* Build a dictionary that store which letter has what frequency
* Check the sequence in the dictionary
    * write bits starting from the root node
    * one byte is written to file
* Write the frequency and compressed data to the output file

2. De-compression

* Read the frequency and compressed data from the file
* Build a tree from the frequencies
* Build a dictionary that store which letter has what frequency
* Check the sequence in the dictionary
    * get letter for the bits
    * get one byte data
* Write the letters to the output file

Execution

1. Clone the repo

$ git clone https://github.com/AP-Atul/Huffman-Compression.git

2. Run the makefile for the serial

$ make serial
or
$ make -B serial

3. Run the binary files on input file to compress

$ ./bin/compress input.txt input.compressed

4. Or decompression

$ ./bin/decompress input.compressed input_decompressed.txt

For parallel with cuda

$ make -B cuda

• Compression

$ ./bin/compress_cuda input.txt input.compressed

4. Decompression (serial)

$ ./bin/decompress input.compressed input_decompressed.txt

Issues

For any help or problems with the code raise an issue.