enable more gpu test (#1042)

* enable more gpu test

* enable more

* more test

* more tests

src/ngraph/runtime/gpu/unit_test.manifest

 abc_int64
-abc_tbb
-aliased_output
-backwards_broadcast0
-backwards_broadcast1
-backwards_concat_axis_0
-backwards_concat_axis_1
-backwards_concat_vector
-backwards_divide
-backwards_dot_tensor2_tensor2
-backwards_dot_tensor3_tensor3
-backwards_log
-backwards_power
-backwards_relu
 backwards_replace_slice
-backwards_reverse_3d_02
 backwards_reverse_sequence_n4d2c3h2w2
 backwards_reverse_sequence_n3_c2_h3
-backwards_select
-backwards_select_nested
-backwards_sign
 backwards_slice
 backwards_softmax_3d
 backwards_softmax_all
 backwards_softmax_axis
 backwards_softmax_underflow
-backwards_subtract
-backwards_sum_m2s
-backwards_sum_m2v_0
-backwards_sum_m2v_1
-backwards_sum_v2s
-backwards_tan
-backwards_tanh
 batch_norm_one_output
 batch_norm_three_outputs
 broadcast_vector_rowwise_int64
-broadcast_vector_rowwise_reversed
 computation_reuse
 concat_matrix_int64
 constant_broadcast
@@ -69,19 +44,10 @@ divide_by_zero_int32
 dot_4d_5d_multi_axis_big_fp64_VERY_SLOW
 dot_matrix_vector_int64
 function_call
-logical_and
-logical_or
 mkldnn_layouts
-negative
-numeric_double_inf
 numeric_double_nan
 numeric_float_inf
 numeric_float_nan
-one_hot_scalar_fp_nonint_in_3
-one_hot_scalar_oob_in_3
-one_hot_vector_1_barely_oob
-one_hot_vector_1_far_oob
-one_hot_vector_1_fp_nonint
 parameter_as_output
 reduce_window_emulating_max_pool_1d_1channel_1image
 reduce_window_emulating_max_pool_1d_1channel_2image
@@ -110,28 +76,3 @@ tensor_constant_float32
 tensor_constant_int64
 tensor_constant_with_op
 tensorview_custom_mem
-zero_sized_abs
-zero_sized_acos
-zero_sized_asin
-zero_sized_atan
-zero_sized_ceiling
-zero_sized_cos
-zero_sized_cosh
-zero_sized_divide
-zero_sized_eq
-zero_sized_exp
-zero_sized_floor
-zero_sized_greater
-zero_sized_greatereq
-zero_sized_less
-zero_sized_lesseq
-zero_sized_log
-zero_sized_not
-zero_sized_not_equal
-zero_sized_power
-zero_sized_sign
-zero_sized_sin
-zero_sized_sinh
-zero_sized_subtract
-zero_sized_tan
-zero_sized_tanh