case 55: /* negf freg, freg */
		o1 = c.oprrr(p, p.As, int(p.Scond))

		rf := int(p.From.Reg)
		rt := int(p.To.Reg)

		o1 |= (uint32(rf)&15)<<0 | (uint32(rt)&15)<<12

	case 56: /* move to FP[CS]R */
		o1 = ((uint32(p.Scond)&C_SCOND)^C_SCOND_XOR)<<28 | 0xee1<<16 | 0xa1<<4

		o1 |= (uint32(p.From.Reg) & 15) << 12

	case 57: /* move from FP[CS]R */

      attributes {tf_saved_model.exported_names = ["a", "b", "c"]} {
    return
  }
}

// -----

module attributes {tf.versions = {bad_consumers = [], min_consumer = 12 : i32, producer = 1512 : i32}} {
  // CHECK-LABEL @lowers_string_ops
  // CHECK-DAG: ml_program.global public @vars.Variable_1([]) : tensor<!tf_type.string>

  func.return %0 : tensor<1x512xf32>
}

// CHECK:       func @mean_4d_no_keepdim([[VAL_0:%.*]]: tensor<1x48x48x512xf32>) -> tensor<1x512xf32> {
// CHECK:           [[VAL_1:%.*]] = "tfl.pseudo_const"(){{.*}}dense<[1, 512]> : tensor<2xi32>

  %result = "tf.TensorArrayGatherV3"(%handle, %indices, %flow_1) {device = "/job:localhost/replica:0/task:0/device:CPU:0", element_shape = #tf_type.shape<1x512>} : (tensor<2x!tf_type.resource<tensor<1x512xf32>>>, tensor<1xi32>, tensor<f32>) -> tensor<1x1x512xf32>
  func.return %result : tensor<1x1x512xf32>

			isErr:         false,
			expectedCurve: elliptic.P256(),
		},
		"ECDSA-P384": {
			key:           ecdsaPrivKeyP384,
			isErr:         false,
			expectedCurve: elliptic.P384(),
		},
		"ECDSA-P512": {
			key:           ecdsaPrivKeyP521,
			isErr:         false,
			expectedCurve: elliptic.P256(),
		},
		"ED25519": {
			key:           ed25519PrivKey,
			isErr:         true,
			expectedCurve: nil,

}

// -----

// CHECK-LABEL: testSliceConcat
func.func @testSliceConcat(%arg0: tensor<3x1x512xf32>) -> tensor<3x1x512xf32> {
  %0 = "mhlo.slice"(%arg0) <{limit_indices = dense<[1, 1, 512]> : tensor<3xi64>, start_indices = dense<[0, 0, 0]> : tensor<3xi64>, strides = dense<1> : tensor<3xi64>}> : (tensor<3x1x512xf32>) -> tensor<1x1x512xf32>

  TensorShape y_shape({1, 5, 2});
  auto y = ExpandDims(scope_, x, 0);
  RunTest(x, x_shape, y, y_shape);
}

TEST_F(ArrayGradTest, SqueezeGrad) {
  TensorShape x_shape({1, 5, 1, 2});
  auto x = Placeholder(scope_, DT_FLOAT, Placeholder::Shape(x_shape));
  TensorShape y_shape({5, 2});
  auto y = Squeeze(scope_, x);
  RunTest(x, x_shape, y, y_shape);
}

	CMPW	R7, R8                 // 1978
	CMPW	R9, $-32768            // a79e8000
	CMPW	R9, $-32769            // c29dffff7fff
	CMPW	R9, $-2147483648       // c29d80000000
	CMPWU	R1, R2                 // 1512
	CMPWU	R3, $4294967295        // c23fffffffff

	TMHH	R1, $65535             // a712ffff
	TMHL	R2, $1                 // a7230001
	TMLH	R3, $0                 // a7300000
	TMLL	R4, $32768             // a7418000

		{"-0", MaxPrec, "-0", Exact},
		{"-Inf", MaxPrec, "-Inf", Exact},
		{"+Inf", MaxPrec, "+Inf", Exact},

		// just a few regular cases - general rounding is tested elsewhere
		{"1.5", 1, "2", Above},
		{"-1.5", 1, "-2", Below},
		{"123", 1e6, "123", Exact},
		{"-123", 1e6, "-123", Exact},
	} {
		x := makeFloat(test.x).SetPrec(test.prec)
		prec := test.prec
		if prec > MaxPrec {
			prec = MaxPrec
		}

DATA ·IEEEConst+1488(SB)/8,$0x000000011b6ddf04
DATA ·IEEEConst+1496(SB)/8,$0x000000013eb95bc0

	/* x^164928 mod p(x), x^164864 mod p(x) */
DATA ·IEEEConst+1504(SB)/8,$0x00000001f3e29ccc
DATA ·IEEEConst+1512(SB)/8,$0x00000001dfc9fdfc

	/* x^163904 mod p(x), x^163840 mod p(x) */
DATA ·IEEEConst+1520(SB)/8,$0x0000000135ae7562
DATA ·IEEEConst+1528(SB)/8,$0x00000000cd028bc2

	/* x^162880 mod p(x), x^162816 mod p(x) */