This repository is codebase for Runtime Systems project. We implement and evaluate runtime and memory metrics for various benchmarks across three python interpreters - CPython3, PyPy3 and Jython.
To setup and replicate our benchmark data, installing the same version of interpreters is necessary. To install, run
''' github clone link_to_the_repository cd path_to_repository chmod x+w setup.sh ./setup.sh '''
All the benchmarks are located in experiment/benchmarks directory. To run all the benchmarks and record runtime, run
''' python3 run_experiments.py '''
This would automatically run all benchmarks across all interpreters and store the runtimes in results.csv in the main directory.
We use memory_profiler which is the standard python memory profiling tool to collect memory usage statistics. Memory profiler has to be installed individually for each interpreters as follows: ''' python3 -m pip install memory_profiler pypy3 -m pip install memory_profiler jython -m pip install memory_profiler '''
For each benchmark, run ''' python3 -m mprof run benchmark.py mprof plot '''
This would produce a matplot with graph showing memory usage against runtime. Note: Plotting the graph requires matplotlib package.s
