Ruby out performs Crystal significantly on this numerical algorithm

Help & Support
15

OK, after I used the correct version, I confirm using

@[Link("gmp")]
lib LibGMP
  fun mpz_powm = __gmpz_powm(rop : MPZ*, base : MPZ*, exp : MPZ*, mod : MPZ*)
end

def powmodgmp(b, e, m)
  y = BigInt.new
  LibGMP.mpz_powm(y, b.to_big_i, e.to_big_i, m.to_big_i)
  y
end

instead of

@[Link("gmp")]
lib LibGMP
  fun mpz_powm_sec = __gmpz_powm_sec(rop : MPZ*, base : MPZ*, exp : MPZ*, mod : MPZ*)
end

def powmodgmp(b, e, m)
  y = BigInt.new
  LibGMP.mpz_powm_sec(y, b.to_big_i, e.to_big_i, m.to_big_i)
  y
end

produce comparable time with Ruby. :slight_smile:

I also just found out you can use the form without the _sec in Benchmark but not with it.
I assume there are other functions you can use w/o the security versions, which would be nice to document so others aren’t subjected to silent poor performance because of not knowing the version they used is not designed for speed.

4 Likes