Speed-up scalar multiplication in the group

[Nashtare]

Currently the scalar multiplication is done through a simple double-and-add algorithm, skipping the first MSB that is always zero.
We would gain from having other, more efficient methods implemented (not necessarily limited to only 1), among which:

• variable based scalar mult
• Shamir's simulteneous scalar mult (Algorithm 3.48, originally from Straus - can be generalized to more than 2 points, and either constant time with fixed window or variable time with sliding window. Could be beneficial in the Schnorr signature for computing h.P + r.G.)
• Pippenger's algorithm

More here:

• http://www1.spms.ntu.edu.sg/~ccrg/documents/chienning-multiexponentiation.pdf
• https://link.springer.com/content/pdf/10.1007%2F3-540-45537-X_13.pdf

[Nashtare]

Point 2 [Shamir's simultaneous scalar mult] is partially implemented (limited to 2 points, with regular bitwise double-and-add) in commit 7cdb2b3.

[Nashtare]

Pippenger's algorithm is implemented in commit 5b0b2ac.
Closing this issue as of now. If other optimizations with respect to scalar multiplication seem of any interest, we'll open another related issue.