This is the implementation of the KLA and CA algorithms for mapping stream programs onto distributed systems. 

- Requirements:
  + GSL - GNU Scientific Library https://www.gnu.org/software/gsl/

- To compile: make
- To run:
    ./partition -gf <graph_file> -tf <target_file> -m <method> -of <output_file>

- Format of graph file:
  + First line: <number of task (let call N)>
  + Next N lines: 
    <task id/name of task 1> <weight/execution time of task 1>
    ...
    <task id/name of task N> <weight/execution time of task N> 
  
  + Next is an NxN matrix where element (i, j) is the weight of stream connecting task i and task j
  
  
- Format of target file:
  + First line: <number of processing element (let call P)>
  + Second line: P elements, each show the capacity/number of cores of each processing element
  + Next is an PxP matrix where element (i, j) is the bandwidth of the channel connecting processing element j and processing element j

- Output file will be formatted as follows:
 + First line: < number of task (let call N)>
 + Next N lines:
    <task id/name of task 1> <index of machine task 1 is mapped to>
    ...
    <task id/name of task N> <index of machine task N is mapped to>

- Method:
  + 1: CA
  + 2: KLA
  + default: CA