Keep Newton actuators opt-in

Preserve each task's configured actuator path instead of forcing the native Newton adapter in every Warp environment. The native path currently makes representative Go2 physics substeps roughly ten times slower, so it remains available for targeted validation without regressing the default frontend.

Author: hujc7

source/isaaclab_experimental/isaaclab_experimental/envs/direct_rl_env_warp.py

@@ -102,10 +102,6 @@ def __init__(self, cfg: DirectRLEnvCfg, render_mode: str | None = None, **kwargs
             RuntimeError: If a simulation context already exists. The environment must always create one
                 since it configures the simulation context and controls the simulation.
         """
-        # Keep the Warp frontend on the Warp-native actuator path. This covers
-        # implicit, PD, delayed, DC, and neural actuator configurations without
-        # crossing through the Torch-based Isaac Lab actuator loop each step.
-        cfg.sim.use_newton_actuators = True
         # check that the config is valid
         cfg.validate()
         # store inputs to class

source/isaaclab_experimental/isaaclab_experimental/envs/manager_based_env_warp.py

@@ -73,10 +73,6 @@ def __init__(self, cfg: ManagerBasedEnvCfg):
             RuntimeError: If a simulation context already exists. The environment must always create one
                 since it configures the simulation context and controls the simulation.
         """
-        # Keep the Warp frontend on the Warp-native actuator path. This covers
-        # implicit, PD, delayed, DC, and neural actuator configurations without
-        # crossing through the Torch-based Isaac Lab actuator loop each step.
-        cfg.sim.use_newton_actuators = True
         # check that the config is valid
         cfg.validate()
         # store inputs to class