Open bennn opened 1 year ago
I think the 2nd official step would be a good option. We could maybe install the package, get the sample benchmarking script, and change it so that it runs the benchmarks and measures the performance. Im not sure how we can implement this or whether it is possible, but we could try.
The real problem is that the benchmarks are Racket code. We'll need to port them to Python.
It's not easy!
To see why, try porting the untyped sieve
code to Python. Very hard.
benchmark idea, from Craig White via email: https://github.com/gerzytet/star-battles-solver
Can we port the Racket GTP Benchmarks to Static Python?
https://docs.racket-lang.org/gtp-benchmarks/index.html
It will be tough, but it'll be good for us and for the GTP suite.