*   Enable JIT-compiled i64-indexed kernels on GPU for large tensors with more than 2**32 elements.
    *   Unary GPU kernels: Abs, Atanh, Acos, Acosh, Asin, Asinh, Atan, Cos, Cosh, Sin, Sinh, Tan, Tanh.
    *   Binary GPU kernels: AddV2, Sub, Div, DivNoNan, Mul, MulNoNan, FloorDiv, Equal, NotEqual, Greater, GreaterEqual, LessEqual, Less.

* `tf.lite`

= u2 - u1 STx(h) LDy(z1in) CALL p256MulInternal<>(SB) // z3 = h * z1 LDP 4*16(a_ptr), (acc0, acc1)// iff select[0] == 0, z3 = z1 LDP 5*16(a_ptr), (acc2, acc3) ANDS $1, hlp1, ZR CSEL EQ, acc0, y0, y0 CSEL EQ, acc1, y1, y1 CSEL EQ, acc2, y2, y2 CSEL EQ, acc3, y3, y3 LDP p256one<>+0x00(SB), (acc0, acc1) LDP p256one<>+0x10(SB), (acc2, acc3) ANDS $2, hlp1, ZR // iff select[1] == 0, z3 = 1 CSEL EQ, acc0, y0, y0 CSEL EQ, acc1, y1, y1 CSEL EQ, acc2, y2, y2 CSEL EQ, acc3, y3, y3 LDx(z1in) MOVD res+0(FP), t0...