parameter's event will be included execution args

- all primitives generate event.
- remove reorder from test-case

src/plugins/intel_gpu/src/graph/primitive_inst.cpp

@@ -909,8 +909,6 @@ event::ptr primitive_inst::execute(const std::vector<event::ptr>& events) {
         if (out_of_order_queue || _impl->is_cpu() || (can_be_optimized() && needs_completion_event() && !is_output())) {
             dependencies.reserve(dependencies.size() + _exec_deps.size());
             for (auto& input : _exec_deps) {
-                if (input->is_input() && !out_of_order_queue)
-                    continue;
                 auto id = input->id();
                 try {
                     // if the requested event does not exists it means that it has not been executed, so the processing_order is

src/plugins/intel_gpu/tests/unit/test_cases/condition_gpu_test.cpp

@@ -367,7 +367,6 @@ class condition_gpu_tests: public ::testing::Test {
         const primitive_id branch_input_id = "branch_input";
         const primitive_id model_input = "input";
         const primitive_id predicate_input = "predicate";
-        const primitive_id reorder_id = "reorder";
         const primitive_id tranpose = "transpose";
 
         cldnn::topology topology;
@@ -391,8 +390,7 @@ class condition_gpu_tests: public ::testing::Test {
         condition::branch branch_true = generate_simple_branch(true, branch_input_id, data_types::f32);
         condition::branch branch_false = generate_simple_branch(false, branch_input_id, data_types::f32);
 
-        topology.add(reorder(reorder_id, input_info(predicate_input), { {d1, -1, -1, d2}, data_types::f32, format::bfyx }));
-        topology.add(condition(condition_id, { reorder_id, tranpose }, branch_true, branch_false));
+        topology.add(condition(condition_id, { input_info(predicate_input), tranpose }, branch_true, branch_false));
 
         tests::random_generator rg(GET_SUITE_NAME);
         std::vector<uint8_t> predicate_data_true = { 1 };