Docs status:
To build the executable both for tests and source code run
make
To clean the .o files in the build directories run
make clean
To build and execute test cases run
make test [NUM_PROCESSES=NUM_PROCESSES]
To generate measures (TAKES A LOT OF TIME! Our measures are already included so you should skip this step) run
make measures
To generate a file with random numbers run
make generate_file [N=N[:10000]] [MAX=MAX[:2147483647]] [MIN=MIN[:-MAX]] [FILENAME=path/to/inputfile[:input/in.txt]]
To build and execute the source code run
make run [NUM_PROCESSES=NUM_PROCESSES] [FILENAME=path/to/inputfile[:input/in.txt]]
To build and execute SILENTLY the source code run
make -s run [NUM_PROCESSES=NUM_PROCESSES] [FILENAME=path/to/inputfile[:input/in.txt]]
To generate a file with random numbers and execute a single run of the mergesort run
make generate_file run [N=N[:10000]] [MAX=MAX[:2147483647]] [MIN=MIN[:-MAX]] [FILENAME=path/to/inputfile[:input/in.txt]]
To execute one of these commands (all dir test_dir clean run test) only for serial or parallel version add "serial" or "parallel" respectively after make.
Example:
make -s serial run [NUM_PROCESSES=NUM_PROCESSES] [FILENAME=path/to/inputfile[:input/in.txt]]
Results can be found in the measures directory, divided by problem size and gcc optimization option.
The speedup curves are generated taking the relative serial execution as reference.