Fractions are slow in PyPy
This is why the Markov decision process in pyperformance is slower in PyPy.
from time import perf_counter from fractions import Fraction def primes(n): # returns first n primes primes =  i=3 while len(primes)<n: if all(i%p for p in primes): primes.append(i) i+=2 return primes T=1000 N=1000 a = [Fraction(i)/Fraction(p) for i,p in enumerate(primes(N))] # Worst case CPython: 3.60 secs | PyPy: 142.1 secs t = perf_counter() for _ in range(T): sum(a) print(perf_counter()-t) a = [Fraction(i)/Fraction(N) for i in range(N)] # Good case CPython: 1.23 secs | PyPy: 0.10 secs t = perf_counter() for _ in range(T): sum(a) print(perf_counter()-t)
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information