I'll answer just #2, since it's the only one I've got anything remotely intelligent to say, but for your Python code, you're creating an intermediate list in is_prime, whereas you're using .map in your all in Ruby which is just iterating.
If you change your is_prime to:
def is_prime(n):
return all((n%j > 0) for j in range(2, n))
they're on par.
I could optimize the Python further, but my Ruby isn't good enough to know when I've given more of an advantage (e.g., using xrange makes Python win on my machine, but I don't remember if the Ruby range you used creates an entire range in memory or not).
EDIT: Without being too silly, making the Python code look like:
import time
def is_prime(n):
return all(n % j for j in xrange(2, n))
def primes_below(x):
return [(j-6, j) for j in xrange(9, x + 1) if is_prime(j) and is_prime(j-6)]
a = int(round(time.time() * 1000))
print(primes_below(10*1000))
b = int(round(time.time() * 1000))
print(str((b-a)) + " mils")
which doesn't change much more, puts it at 1.5s for me, and, with being extra silly, running it with PyPy puts it at .3s for 10K, and 21s for 100K.
