better // 2) CSEL might not be constant time on all ARM processors LDP 0*16(a_ptr), (R4, R5) LDP 1*16(a_ptr), (R6, R7) LDP 2*16(a_ptr), (R8, R9) LDP 0*16(b_ptr), (R16, R17) LDP 1*16(b_ptr), (R19, R20) LDP 2*16(b_ptr), (R21, R22) CSEL EQ, R16, R4, R4 CSEL EQ, R17, R5, R5 CSEL EQ, R19, R6, R6 CSEL EQ, R20, R7, R7 CSEL EQ, R21, R8, R8 CSEL EQ, R22, R9, R9 STP (R4, R5), 0*16(res_ptr) STP (R6, R7), 1*16(res_ptr) STP (R8, R9), 2*16(res_ptr) LDP 3*16(a_ptr), (R4, R5) LDP 4*16(a_ptr), (R6, R7) LDP 5*16(a_ptr),...