flen := f[start:start+len], f[start+len:start+len+len] for j := 0; j < len; j += 2 { t := fieldMontgomeryMul(zeta, flen[j]) flen[j] = fieldSub(f[j], t) f[j] = fieldAdd(f[j], t) // Unroll by 2 for performance. t = fieldMontgomeryMul(zeta, flen[j+1]) flen[j+1] = fieldSub(f[j+1], t) f[j+1] = fieldAdd(f[j+1], t) } } } // Unroll len = 4, 2, and 1. for start := 0; start < 256; start += 8 { m++ zeta := zetas[m] t := fieldMontgomeryMul(zeta, f[start+4]) f[start+4] = fieldSub(f[start], t) f[start] = fieldAdd(f[start],...