VPERMB -7(CX)(DX*8), Y31, K7, Y22                  // 62e205278db4d1f9ffffff
	VPERMB Y5, Y5, K7, Y22                             // 62e2552f8df5
	VPERMB Y19, Y5, K7, Y22                            // 62a2552f8df3
	VPERMB Y31, Y5, K7, Y22                            // 6282552f8df7
	VPERMB 17(SP)(BP*1), Y5, K7, Y22                   // 62e2552f8db42c11000000

	VPOPCNTQ 15(DX), K4, X23                           // 62e2fd0c55ba0f000000
	VPOPCNTQ X24, K4, X11                              // 6212fd0c55d8
	VPOPCNTQ X14, K4, X11                              // 6252fd0c55de
	VPOPCNTQ X0, K4, X11                               // 6272fd0c55d8
	VPOPCNTQ 99(R15)(R15*4), K4, X11                   // 6212fd0c559cbf63000000

		{"VADDPD.SAE.SAE X0, X1, X2", `duplicate suffix "SAE"`},
		{"VADDPD.RZ_SAE.SAE X0, X1, X2", `bad suffix combination`},

		// BSWAP on 16-bit registers is undefined. See #29167,
		{"BSWAPW DX", `unrecognized instruction`},
		{"BSWAPW R11", `unrecognized instruction`},
	})
}

func testBadInstParser(t *testing.T, goarch string, tests []badInstTest) {
	for i, test := range tests {

// i.e., d(y_1 + y_2 + ...)/dx_1, d(y_1 + y_2 + ...)/dx_2...
//
// `dx` are used as initial gradients (which represent the symbolic partial
// derivatives of some loss function `L` w.r.t. `y`).
// `dx` must be nullptr or have size `ny`.
// If `dx` is nullptr, the implementation will use dx of `OnesLike` for all
// shapes in `y`.

	PINSRD	$1, (AX), X0
	PINSRD	$2, foo+4(FP), X0

// Was bug: LOOP is a branch instruction.
	JCS	2(PC)
loop:
	LOOP	loop // LOOP

// Tests for TLS reference.
	MOVL    (TLS), AX
	MOVL    8(TLS), DX

// LTYPE0 nonnon	{ outcode(int($1), &$2); }
	RET

	{"(AX)(CX*8)", "(AX)(CX*8)"},
	{"(BP)(CX*4)", "(BP)(CX*4)"},
	{"(BP)(DX*4)", "(BP)(DX*4)"},
	{"(BP)(R8*4)", "(BP)(R8*4)"},
	{"(BX)", "(BX)"},
	{"(DI)", "(DI)"},
	{"(DI)(BX*1)", "(DI)(BX*1)"},
	{"(DX)", "(DX)"},
	{"(R9)", "(R9)"},
	{"(R9)(BX*8)", "(R9)(BX*8)"},
	{"(SI)", "(SI)"},
	{"(SI)(BX*1)", "(SI)(BX*1)"},
	{"(SI)(DX*1)", "(SI)(DX*1)"},
	{"(SP)", "(SP)"},
	{"(SP)(AX*4)", "(SP)(AX*4)"},

     *
     *    dX = U / Y
     *    dY = -U*X / Y^2 = (X/Y) * -U / Y = -U*Z / Y
     *
     */

    AbstractTensorHandle* upstream_grad = grad_outputs[0];
    AbstractTensorHandle* Y = forward_inputs_[1];
    AbstractTensorHandle* Z = forward_outputs_[0];

    // Calculate dX =  U / Y
    std::string name = "Div_Grad_X";
    TF_RETURN_IF_ERROR(

	VPGATHERDQ X2, 664(BX*1), X1    // ERROR "invalid instruction"
	VPGATHERDQ Y2, (BP)(AX*2), Y1   // ERROR "invalid instruction"
	VPGATHERDQ Y5, 664(DX*8), Y6    // ERROR "invalid instruction"
	VPGATHERDQ Y5, (DX), Y0         // ERROR "invalid instruction"
	// VM/X rejects Y index register.
	VPGATHERDQ Y5, 664(Y14*8), Y6   // ERROR "invalid instruction"

    ASSERT_EQ(errors::OK, s.code()) << s.message();
    y.reset(y_raw);
  }

  float expected_dx[1] = {7.0f};
  ASSERT_NO_FATAL_FAILURE(CompareNumericalAndManualGradients(
      MulModel, ctx_.get(), {x.get(), y.get()}, 0, expected_dx, 1,
      UseFunction()));
}

#ifdef PLATFORM_GOOGLE
INSTANTIATE_TEST_SUITE_P(
    UnifiedCAPI, GradientCheckerTest,

<mxfile host="65bd71144e">
    <diagram id="jyERGzDynktFHFRGN0ph" name="Page-1">
        <mxGraphModel dx="2738" dy="2173" grid="1" gridSize="10" guides="1" tooltips="1" connect="1" arrows="1" fold="1" page="1" pageScale="1" pageWidth="1920" pageHeight="1200" math="0" shadow="0">
            <root>
                <mxCell id="0"/>
                <mxCell id="1" parent="0"/>