The cost of solve (and verify) is dominated byBigInt multiplications. At some size of modulus, an FFT-based multiplication (i.e.. Schönhage–Strassen) will be faster than the direct method, and also can potentially be parallelised (undermining the desired non-parallelisableness of the challenge).
It would be good to know what modulus this is.
The cost of solve (and verify) is dominated byBigInt multiplications. At some size of modulus, an FFT-based multiplication (i.e.. Schönhage–Strassen) will be faster than the direct method, and also can potentially be parallelised (undermining the desired non-parallelisableness of the challenge).
It would be good to know what modulus this is.