After we have good data running these benchmarks through the default Python + retic, we should try running them again through PyPy and see if that changes the slowdown due to gradual typing.
Just running them with the different backend should be all the "work" we have to do.
http://pypy.org/
After we have good data running these benchmarks through the default Python + retic, we should try running them again through PyPy and see if that changes the slowdown due to gradual typing.
Just running them with the different backend should be all the "work" we have to do.
http://pypy.org/