feat: double method for FieldElement

[tcoratger]

Description

This pull request proposes a performance enhancement in the implementation of elliptic curve operations. Instead of using a full multiplication operation like FieldElement::TWO * a to handle doubling (2 * a), it suggests utilizing the double method available in the FieldElement struct.

The rationale behind this enhancement is that the doubling operation (a + a) requires only one addition between field elements, which is more performant compared to a full multiplication. Therefore, replacing multiplication with addition can improve the efficiency of various computations involving point doubling in elliptic curve arithmetic.

Changes Made

• Added logic to utilize the double method from the FieldElement struct for doubling operations.
• Updated relevant computations to use the double method where appropriate.

[xJonathanLEI]

Ran the benchmark on this PR branch against master:

class_hash              time:   [17.715 ms 17.755 ms 17.804 ms]
                        change: [+0.0908% +0.6897% +1.1519%] (p = 0.01 < 0.05)

ecdsa_get_public_key    time:   [91.500 µs 91.624 µs 91.803 µs]
                        change: [-12.552% -11.651% -11.003%] (p = 0.00 < 0.05)

ecdsa_recover           time:   [376.48 µs 378.87 µs 381.03 µs]
                        change: [-4.9179% -4.3248% -3.7237%] (p = 0.00 < 0.05)

ecdsa_sign              time:   [136.55 µs 136.85 µs 137.24 µs]
                        change: [-11.674% -11.152% -10.675%] (p = 0.00 < 0.05)

ecdsa_verify            time:   [389.95 µs 390.62 µs 391.22 µs]
                        change: [-0.6257% -0.2373% +0.1272%] (p = 0.23 > 0.05)

pedersen_hash           time:   [24.607 µs 24.649 µs 24.700 µs]
                        change: [-0.6410% +0.2009% +1.1220%] (p = 0.66 > 0.05)

poseidon_hash           time:   [7.9407 µs 7.9519 µs 7.9660 µs]
                        change: [-15.201% -14.834% -14.310%] (p = 0.00 < 0.05)

poseidon_hash_single    time:   [7.5840 µs 7.6086 µs 7.6400 µs]
                        change: [-17.138% -16.698% -16.290%] (p = 0.00 < 0.05)

poseidon_hash_many      time:   [15.201 µs 15.218 µs 15.239 µs]
                        change: [-19.208% -19.062% -18.923%] (p = 0.00 < 0.05)

rfc6979_generate_k      time:   [1.6403 µs 1.6422 µs 1.6459 µs]
                        change: [+8.1256% +8.6450% +9.3945%] (p = 0.00 < 0.05)

So we got some pretty decent improvements on ECDSA stuff, and even more impressive ones on Poseidon. Thank you!

[xJonathanLEI]

LGTM