source 2002 Izu--Takagi "A fast parallel elliptic curve multiplication resistant against side channel attacks", page 296, Formula 3
compute T1 = X1^2
compute T2 = Z1^2
compute T3 = a T2
compute T4 = T1 - T3
compute T5 = T4^2
compute T6 = b T2
compute T7 = X1 Z1
compute T8 = T6 T7
compute T9 = 8 T8
compute X3 = T5 - T9
compute T10 = T1 + T3
compute T11 = T7 T10
compute T12 = T6 T2
compute T13 = T11 + T12
compute Z3 = 4 T13