The archive of the source code can be downloaded here : PLI.tar.gz

After extracting the archive, the binary executable "pli" can be generated in 2 ways :

• make pli

• make plig : generates an executable that uses the gmp arithmetic multi-precision library (has to be installed before of course)

Program usage :

./pli -m selected_model -i input_trace-file -o output_file 
[-e admissible_error_level_per_phase]
[-p maximum_interpolation_periode_per_phase]
[-s] 
where all options in [ ] are optional and -s sorts phases in a chronological order in the output file
selected_model = af, df, ar, dr, av or dv
af = adjacent intervals of constant size
df = distant intervals of constant size
ar = equal adjacent intervals of constant size (repetitive patterns)
dr = equal distant intervals of constant size (repetitive patterns)
av = adjacent intervals of variable sizes
dv = distant intervals of variable sizes

The input trace file has to be an ascii file where all values are seperated by a space. See the file trace.txt as an example. See also the file result.txt showing the resulting file obtained with trace.txt as input and with the "af" model.

Questions and comments can be sended to Philippe Clauss (clauss_at_icps.u-strasbg.fr)