Point Addition Performance

[iceland2k14]

Your library performance is Very good for Scalar Multiplication. But i am trying for Point Addition and getting slower speed.

pub1 = bit.Key.from_int(1)._pk.public_key
timeit pub1.combine_keys([pub1, pub1])
8.81 µs ± 287 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

In comparison, by using gmpy2 with https://github.com/iceland2k14/bsgs/blob/main/gmp_ec.py
timeit ec.G + ec.G
6.56 µs ± 106 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

Maybe i am not trying in proper way to do Point Addition with bit or coincurve. Can you please show ?

[ofek]

Can you give me a way to test like this:

$ python -m timeit -s "from coincurve import PrivateKey,PublicKey;pub_key=PrivateKey.from_int(1).public_key;keys=[pub_key, pub_key]" "PublicKey.combine_keys(keys)"
50000 loops, best of 5: 5.87 usec per loop

[iceland2k14]

Just to avoid Point Doubling inside, Adding G + 2G, here they are :

python -m timeit -s "from coincurve import PrivateKey,PublicKey;pub_key2=PrivateKey.from_int(2).public_key;pub_key1=PrivateKey.from_int(1).public_key;keys=[pub_key1, pub_key2]" "PublicKey.combine_keys(keys)"
50000 loops, best of 5: 8.41 usec per loop

python -m timeit -s "import gmp_ec as ec;pub_key1=ec.G;pub_key2=ec.Scalar_Multiplication(2)" "ec.Point_Addition(pub_key1, pub_key2)"
50000 loops, best of 5: 3.9 usec per loop