fix bug (#49081)

python/paddle/distributed/fleet/meta_optimizers/graph_execution_optimizer.py

@@ -71,12 +71,6 @@ class GraphExecutionOptimizer(MetaOptimizerBase):
         if trainer_id == 0 and not paddle.is_compiled_with_npu():
             wait_server_ready(other_trainers)
 
-        if (
-            build_strategy.reduce_strategy
-            == BuildStrategy.ReduceStrategy._NoReduce
-        ):
-            return
-
         if core.is_compiled_with_cuda():
             comm_id_var = startup_program.global_block().create_var(
                 name="NCCLID", persistable=True, type=core.VarDesc.VarType.RAW