diff --git a/Scheduler/scheduler.py b/Scheduler/scheduler.py
index 834698a6..10358321 100644
--- a/Scheduler/scheduler.py
+++ b/Scheduler/scheduler.py
@@ -140,7 +140,7 @@ def __str__(self):
     def register_model(model_name : str, compiled_model):
         SchedulerDNNModel.MODEL_MAP[model_name] = compiled_model
 
-class ExecutionEngine:
+class PyTorchSimRunner:
     PARTITION_BUSY = 0
     PARTITION_IDLE = 1
     SELECT_NOTHING = 2
@@ -262,7 +262,7 @@ def launch_kernel(self, current_cycle, partion_idx=0):
         self.partition_state[partion_idx] = self.PARTITION_BUSY
         return self.backend_simulator.launch(onnx_path, attribute_path, current_cycle, partion_idx)
 
-class FIFOExecutionEngine(ExecutionEngine):
+class FIFORunner(PyTorchSimRunner):
     def __init__(self, backend_simulator: BackendSimulator, num_partion=1) -> None:
         super().__init__(backend_simulator, num_partion)
 
@@ -297,7 +297,7 @@ def select_kernel(self, partition_idx):
         # No proper kernel now
         return self.SELECT_NOTHING
 
-class RRExecutionEngine(ExecutionEngine):
+class RoundRobinRunner(PyTorchSimRunner):
     def __init__(self, backend_simulator: BackendSimulator, num_partion=1) -> None:
         super().__init__(backend_simulator, num_partion)
         self.next_pointer = None
@@ -360,9 +360,9 @@ def __init__(self, num_request_queue=1, max_batch=1, engine_select=FIFO_ENGINE,
         self.backend_simulator = BackendSimulator(backend_path, backend_config)
         self.backend_simulator.interactive_simulation()
         if engine_select == Scheduler.FIFO_ENGINE:
-            self.execution_engine = FIFOExecutionEngine(self.backend_simulator, self.num_request_queue)
+            self.execution_engine = FIFORunner(self.backend_simulator, self.num_request_queue)
         elif engine_select == Scheduler.RR_ENGINE:
-            self.execution_engine = RRExecutionEngine(self.backend_simulator, self.num_request_queue)
+            self.execution_engine = RoundRobinRunner(self.backend_simulator, self.num_request_queue)
         else:
             print(f"Not supporetd engine type {engine_select}")
             exit(1)
@@ -480,7 +480,7 @@ def run(self, until_time):
         def execute_cycle():
             launch_ret_info = []
             for i in range(self.execution_engine.num_partion):
-                if self.execution_engine.partition_state[i] == ExecutionEngine.PARTITION_IDLE:
+                if self.execution_engine.partition_state[i] == PyTorchSimRunner.PARTITION_IDLE:
                     ret = self.execution_engine.launch_kernel(self.current_cycle, i)
                     launch_ret_info.append(ret)
 
@@ -495,7 +495,7 @@ def execute_cycle():
 
             for core_idx in result_list:
                 # Kernel is finished. So set idle state
-                self.execution_engine.partition_state[core_idx] = ExecutionEngine.PARTITION_IDLE
+                self.execution_engine.partition_state[core_idx] = PyTorchSimRunner.PARTITION_IDLE
 
             return result_list
 
@@ -543,4 +543,4 @@ def msec_to_cycle(self, msec):
             return msec
 
         freq = self.backend_simulator.get_core_freq()
-        return int(msec * (freq / 1000))
\ No newline at end of file
+        return int(msec * (freq / 1000))
diff --git a/experiments/gemm.py b/experiments/gemm.py
index a1fdcff6..e7a639ad 100644
--- a/experiments/gemm.py
+++ b/experiments/gemm.py
@@ -48,7 +48,7 @@ def custom_matmul(a, b):
     if 'BACKENDSIM_SPIKE_ONLY' in os.environ:
         del os.environ['BACKENDSIM_SPIKE_ONLY']
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     run_matmul(size[0], size[1], size[2], config)
diff --git a/scripts/ILS_experiment/test_matmul.py b/scripts/ILS_experiment/test_matmul.py
index 09cc407d..667dfc66 100644
--- a/scripts/ILS_experiment/test_matmul.py
+++ b/scripts/ILS_experiment/test_matmul.py
@@ -60,7 +60,7 @@ def custom_matmul(bias, a, b):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_matmul(device, *shape)
diff --git a/scripts/chiplet_prep.py b/scripts/chiplet_prep.py
index 168532f1..32f7ad50 100644
--- a/scripts/chiplet_prep.py
+++ b/scripts/chiplet_prep.py
@@ -61,8 +61,8 @@ def modify_file(dump_path, name, address_numa_stride=None, subgraph_map=None):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     parser = argparse.ArgumentParser(description='Process folder argument.')
     parser.add_argument('size', type=int, help='Folder value', default=256)
diff --git a/test_extension_backend.py b/test_extension_backend.py
index f0a9353a..5e6427ef 100644
--- a/test_extension_backend.py
+++ b/test_extension_backend.py
@@ -22,8 +22,8 @@
 from tests.Fusion.test_matmul_activation import test_matmul_activation
 
 if __name__ == "__main__":
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     #test_vectoradd(device, (47, 10))
     #test_vector_scalar_add(device, (10, 10))
diff --git a/tests/Diffusion/test_diffusion.py b/tests/Diffusion/test_diffusion.py
index 03d1b721..c5170209 100644
--- a/tests/Diffusion/test_diffusion.py
+++ b/tests/Diffusion/test_diffusion.py
@@ -553,8 +553,8 @@ def test_upsample2d(
     args = parser.parse_args()
 
     sys.path.append(os.environ.get("TORCHSIM_DIR", "/workspace/PyTorchSim"))
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
 
     #test_upsample2d(device)
diff --git a/tests/Fusion/test_addmm_residual.py b/tests/Fusion/test_addmm_residual.py
index a5e05182..ef753a67 100644
--- a/tests/Fusion/test_addmm_residual.py
+++ b/tests/Fusion/test_addmm_residual.py
@@ -43,8 +43,8 @@ def addmm_residual(a, b, c, d):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_addmm_residual(device, 32, 32, 32)
     test_addmm_residual(device, 128, 128, 128)
diff --git a/tests/Fusion/test_attention_fusion.py b/tests/Fusion/test_attention_fusion.py
index 95bdf165..123376d1 100644
--- a/tests/Fusion/test_attention_fusion.py
+++ b/tests/Fusion/test_attention_fusion.py
@@ -75,8 +75,8 @@ def test_MHA(device, num_heads=12, embed_dim=768, input_seq=512):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_MHA(device)
     # test_Attention(device, head=16, seq=512, d_k=64)
diff --git a/tests/Fusion/test_bmm_reduction.py b/tests/Fusion/test_bmm_reduction.py
index 42e38095..4f4d3ad6 100644
--- a/tests/Fusion/test_bmm_reduction.py
+++ b/tests/Fusion/test_bmm_reduction.py
@@ -42,8 +42,8 @@ def bmm(a, b):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     #test_bmm_reduce(device)
     test_bmm_reduce(device, 12, 512)
diff --git a/tests/Fusion/test_conv_fusion.py b/tests/Fusion/test_conv_fusion.py
index 42210b13..694f3bb9 100644
--- a/tests/Fusion/test_conv_fusion.py
+++ b/tests/Fusion/test_conv_fusion.py
@@ -101,8 +101,8 @@ def custom_conv_bn_relu(a, b, bias, c, d, e, f):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
 
     # Vanila test
diff --git a/tests/Fusion/test_matmul_activation.py b/tests/Fusion/test_matmul_activation.py
index 2381bd8c..2f1d014f 100644
--- a/tests/Fusion/test_matmul_activation.py
+++ b/tests/Fusion/test_matmul_activation.py
@@ -73,8 +73,8 @@ def test_matmul_activation(device, batch_size=16, input_size=32, output_size=8,
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_matmul_activation(device)
     test_matmul_activation(device, batch_size=32, input_size=32, output_size=32, activation_fn="sigmoid")
diff --git a/tests/Fusion/test_matmul_reduction.py b/tests/Fusion/test_matmul_reduction.py
index 31ea1b0d..df8cf969 100644
--- a/tests/Fusion/test_matmul_reduction.py
+++ b/tests/Fusion/test_matmul_reduction.py
@@ -89,8 +89,8 @@ def matmul_fused(a, b, c, d):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_matmul_reduce(device, 3072, 512, 768)
     test_matmul_var_mean(device)
diff --git a/tests/Fusion/test_matmul_scalar.py b/tests/Fusion/test_matmul_scalar.py
index 0dcb54f9..0815bb90 100644
--- a/tests/Fusion/test_matmul_scalar.py
+++ b/tests/Fusion/test_matmul_scalar.py
@@ -39,7 +39,7 @@ def matmul_fused(a, b, c):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_matmul_scalar(device)
diff --git a/tests/Fusion/test_prologue_fusion.py b/tests/Fusion/test_prologue_fusion.py
index 797f9e76..b27312a9 100644
--- a/tests/Fusion/test_prologue_fusion.py
+++ b/tests/Fusion/test_prologue_fusion.py
@@ -88,8 +88,8 @@ def bmm(a, b, c, d):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_elem_broadcast_fusion(device)
     test_elem_fusion(device)
diff --git a/tests/Fusion/test_transformer_fusion.py b/tests/Fusion/test_transformer_fusion.py
index 0e500b5b..b1cceb2c 100644
--- a/tests/Fusion/test_transformer_fusion.py
+++ b/tests/Fusion/test_transformer_fusion.py
@@ -203,8 +203,8 @@ def test_EncoderBlock_validation(head=12, embed_dim=768, input_seq=512):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     #test_MHA(device)
     test_EncoderBlock(device)
diff --git a/tests/MLP/test_mlp.py b/tests/MLP/test_mlp.py
index 6f6c9444..31bcefdf 100644
--- a/tests/MLP/test_mlp.py
+++ b/tests/MLP/test_mlp.py
@@ -281,9 +281,9 @@ def train(model, device):
     return
 
 if __name__ == "__main__":
-    from Scheduler.scheduler import ExecutionEngine
+    from Scheduler.scheduler import PyTorchSimRunner
     torch.set_printoptions(threshold=float('inf'), linewidth=600)
-    module = ExecutionEngine.setup_device()
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
 
     test_mlp(device)
diff --git a/tests/MLP/test_mlp_cpu.py b/tests/MLP/test_mlp_cpu.py
index 49f44650..112f5d07 100644
--- a/tests/MLP/test_mlp_cpu.py
+++ b/tests/MLP/test_mlp_cpu.py
@@ -399,7 +399,6 @@ def train(model, device):
 
 
 if __name__ == "__main__":
-    from Scheduler.scheduler import ExecutionEngine
     # torch.set_printoptions(threshold=float('inf'), linewidth=600)
 
     device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
diff --git a/tests/Mixtral_8x7B/test_attention.py b/tests/Mixtral_8x7B/test_attention.py
index aa1af651..6a7747f7 100644
--- a/tests/Mixtral_8x7B/test_attention.py
+++ b/tests/Mixtral_8x7B/test_attention.py
@@ -163,8 +163,8 @@ def test_rmsnorm(device, seq=32):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_rmsnorm(device, seq=1)
     test_concat(device, size1=(1, 8, 64, 64), size2=(1,8,1,64), dim=2)
diff --git a/tests/MoE/test_moe.py b/tests/MoE/test_moe.py
index c5ab8107..ae16f0b0 100644
--- a/tests/MoE/test_moe.py
+++ b/tests/MoE/test_moe.py
@@ -783,9 +783,9 @@ def evaluation(model, evaluation_loader):
         train(opt_model, train_loader)
 
 if __name__ == "__main__":
-    from Scheduler.scheduler import ExecutionEngine
+    from Scheduler.scheduler import PyTorchSimRunner
     torch.set_printoptions(threshold=float('inf'), linewidth=600)
-    module = ExecutionEngine.setup_device()
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
 
     test_moe(device)
diff --git a/tests/test_activation.py b/tests/test_activation.py
index de3542c3..575fc7e8 100644
--- a/tests/test_activation.py
+++ b/tests/test_activation.py
@@ -88,8 +88,8 @@ def test_SwiGLU(device, size=(128, 128)):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_ReLU(device, (47, 10))
     test_ReLU(device, (128, 128))
diff --git a/tests/test_add.py b/tests/test_add.py
index 5e1ab15e..118632d5 100644
--- a/tests/test_add.py
+++ b/tests/test_add.py
@@ -58,8 +58,8 @@ def vectoradd(a, b):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_vectoradd(device, (1, 1))
     test_vectoradd(device, (47, 10))
diff --git a/tests/test_batchnorm.py b/tests/test_batchnorm.py
index f7abacf5..251805f5 100644
--- a/tests/test_batchnorm.py
+++ b/tests/test_batchnorm.py
@@ -37,8 +37,8 @@ def test_BatchNorm(device, size=(1, 16, 64, 64)):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_BatchNorm(device)
     test_BatchNorm(device, size=(1,64, 32, 32))
diff --git a/tests/test_bmm.py b/tests/test_bmm.py
index 6d9279aa..d90410db 100644
--- a/tests/test_bmm.py
+++ b/tests/test_bmm.py
@@ -46,8 +46,8 @@ def bmm(a, b, bias):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_BMM(device)
     test_BMM(device, 2, 256, 128, 256)
diff --git a/tests/test_cnn.py b/tests/test_cnn.py
index aaad2836..54225747 100644
--- a/tests/test_cnn.py
+++ b/tests/test_cnn.py
@@ -53,7 +53,7 @@ def test_CNN(device):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_CNN(device)
diff --git a/tests/test_conv2d.py b/tests/test_conv2d.py
index 21bbfec7..e964319d 100644
--- a/tests/test_conv2d.py
+++ b/tests/test_conv2d.py
@@ -40,8 +40,8 @@ def custom_conv2d(a, b, bias):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     torch._dynamo.config.cache_size_limit = 64
     test_conv2d(device, batch_size=8, in_channels=3, out_channels=32, input_size=32, kernel_size=1, stride=1, padding=0)
diff --git a/tests/test_exponent.py b/tests/test_exponent.py
index c95823cb..e60f8407 100644
--- a/tests/test_exponent.py
+++ b/tests/test_exponent.py
@@ -31,7 +31,7 @@ def exponent(a):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_exponent(device, size=(32, 32))
diff --git a/tests/test_indirect_access.py b/tests/test_indirect_access.py
index b7b20074..c6afaf86 100644
--- a/tests/test_indirect_access.py
+++ b/tests/test_indirect_access.py
@@ -48,8 +48,8 @@ def test_embedding(device, vocab_size, dim):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_indirect_vectoradd(device)
     #test_embedding(device, 1024, 2048)
\ No newline at end of file
diff --git a/tests/test_layernorm.py b/tests/test_layernorm.py
index 1cea9d9f..28e38d37 100644
--- a/tests/test_layernorm.py
+++ b/tests/test_layernorm.py
@@ -41,8 +41,8 @@ def test_LayerNorm(device, size=(64, 64)):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     #test_LayerNorm(device)
     test_LayerNorm(device, shape)
diff --git a/tests/test_matmul.py b/tests/test_matmul.py
index 6f41468b..cd30bd30 100644
--- a/tests/test_matmul.py
+++ b/tests/test_matmul.py
@@ -94,8 +94,8 @@ def custom_linear(a, b, bias):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_matmul(device, 32, 32, 32)
     test_matmul(device, 128, 128, 128)
diff --git a/tests/test_mlp.py b/tests/test_mlp.py
index b8118aa3..423d6e8e 100644
--- a/tests/test_mlp.py
+++ b/tests/test_mlp.py
@@ -109,8 +109,8 @@ def test_optimizer(device):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_mlp(device)
     test_mlp_inf(device, batch_size=1, input_size=256, hidden_size=512, output_size=256)
diff --git a/tests/test_pool.py b/tests/test_pool.py
index 304a5e7c..f5505dba 100644
--- a/tests/test_pool.py
+++ b/tests/test_pool.py
@@ -47,8 +47,8 @@ def avgpool(a):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     #test_maxpool(device, b=1, c=8, h=16, w=16)
     #test_maxpool(device, b=1, c=8, h=112, w=112)
diff --git a/tests/test_reduce.py b/tests/test_reduce.py
index e1a84b7f..4781112d 100644
--- a/tests/test_reduce.py
+++ b/tests/test_reduce.py
@@ -47,8 +47,8 @@ def reduce_sum(a, dim, keepdim):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_reduce_sum(device, (29, 47), 1, keepdim=True)
     test_reduce_sum(device, (17, 68), 0, keepdim=True)
diff --git a/tests/test_resnet.py b/tests/test_resnet.py
index 97c60528..c83f13ba 100644
--- a/tests/test_resnet.py
+++ b/tests/test_resnet.py
@@ -49,7 +49,7 @@ def test_resnet(device, batch=1, model_type='resnet18'):
     args = args.parse_args()
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_resnet(device, model_type=args.model_type)
diff --git a/tests/test_single_perceptron.py b/tests/test_single_perceptron.py
index c7fdca06..beab1c54 100644
--- a/tests/test_single_perceptron.py
+++ b/tests/test_single_perceptron.py
@@ -82,7 +82,7 @@ def weight_update(a, b, lr):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_single_perceptron(device)
diff --git a/tests/test_softmax.py b/tests/test_softmax.py
index 9fba41dd..e6e8cc1e 100644
--- a/tests/test_softmax.py
+++ b/tests/test_softmax.py
@@ -58,8 +58,8 @@ def test_softmax(device, size=(128, 128), dim=1):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_softmax(device, size=(64, 128))
     test_softmax(device, size=(64, 128), dim=0)
diff --git a/tests/test_sparse_core.py b/tests/test_sparse_core.py
index b2b16818..72eda0c8 100644
--- a/tests/test_sparse_core.py
+++ b/tests/test_sparse_core.py
@@ -80,9 +80,9 @@ def test_sparse_mlp(device, batch_size=32, input_size=128, hidden_size=128, outp
     import os
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/root/workspace/PyTorchSim'))
-    from Scheduler.scheduler import ExecutionEngine
+    from Scheduler.scheduler import PyTorchSimRunner
 
-    module = ExecutionEngine.setup_device()
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_sparse_mlp(device, batch_size=8, input_size=16, hidden_size=32, output_size=64)
     
diff --git a/tests/test_sparsity.py b/tests/test_sparsity.py
index 3e079f83..a2493673 100644
--- a/tests/test_sparsity.py
+++ b/tests/test_sparsity.py
@@ -96,8 +96,8 @@ def test_mlp_inf(device, batch_size=64, input_size=64, hidden_size=32, output_si
     )
     args = parser.parse_args()
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
 
     #test_dec_inf(device, sparsity=args.sparsity, block=args.block)
diff --git a/tests/test_stonne.py b/tests/test_stonne.py
index 5e4fe5fb..04ad05a8 100644
--- a/tests/test_stonne.py
+++ b/tests/test_stonne.py
@@ -54,7 +54,7 @@ def test_sparse_mm(device, input_size=128, hidden_size=128, output_size=128, spa
     args = parser.parse_args()
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/root/workspace/PyTorchSim'))
  
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_sparse_mm(device, args.sz, args.sz, args.sz, args.sparsity)
\ No newline at end of file
diff --git a/tests/test_topk.py b/tests/test_topk.py
new file mode 100644
index 00000000..0d5c08ec
--- /dev/null
+++ b/tests/test_topk.py
@@ -0,0 +1,54 @@
+import torch
+import torch._dynamo
+import torch.utils.cpp_extension
+
+def test_result(name, out, cpu_out, rtol=1e-4, atol=1e-4):
+    if torch.allclose(out.cpu(), cpu_out, rtol=rtol, atol=atol):
+        message = f"|{name} Test Passed|"
+        print("-" * len(message))
+        print(message)
+        print("-" * len(message))
+    else:
+        message = f"|{name} Test Failed|"
+        print("-" * len(message))
+        print(message)
+        print("-" * len(message))
+        print("custom out: ", out.cpu())
+        print("cpu out: ", cpu_out)
+        exit(1)
+
+def test_topk(device, size=(128, 128), k=5, dim=-1, largest=True, sorted=True):
+    # dim 해석을 위해 양수 인덱스로 변환
+    dim_ = dim if dim >= 0 else (len(size) + dim)
+    assert 0 <= dim_ < len(size), "dim이 텐서 차원 범위를 벗어났습니다."
+    assert k <= size[dim_], f"k(={k})는 size[dim](={size[dim_]}) 이하여야 합니다."
+
+    def topk_fn(a):
+        return torch.topk(a, k, dim=dim, largest=largest, sorted=sorted)
+
+    x = torch.randn(size)
+    x = x.to(device=device)
+
+    opt_topk = torch.compile(dynamic=False)(topk_fn)
+    res_values, res_indices = opt_topk(x)
+
+    ref_values, ref_indices = torch.topk(x.cpu(), k, dim=dim, largest=largest, sorted=sorted)
+
+    test_result("TopK/values", res_values, ref_values)
+    test_result("TopK/indices", res_indices, ref_indices)
+
+if __name__ == "__main__":
+    import os
+    import sys
+    import argparse
+    sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
+
+    parser = argparse.ArgumentParser(description="Run LayerNorm test with dynamic shape")
+    parser.add_argument('--shape', type=str, default="(512,768)")
+    args = parser.parse_args()
+    shape = tuple(map(int, args.shape.strip('()').split(',')))
+
+    from Scheduler.scheduler import ExecutionEngine
+    module = ExecutionEngine.setup_device()
+    device = module.custom_device()
+    test_topk(device, (128, 128), k=2, dim=-1)
\ No newline at end of file
diff --git a/tests/test_transcendental.py b/tests/test_transcendental.py
index 5f296581..38c2f4f6 100644
--- a/tests/test_transcendental.py
+++ b/tests/test_transcendental.py
@@ -73,8 +73,8 @@ def cos(a):
     args = parser.parse_args()
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_tanh(device)
     test_exp(device)
diff --git a/tests/test_transformer.py b/tests/test_transformer.py
index 4d45707e..a3ac55d7 100644
--- a/tests/test_transformer.py
+++ b/tests/test_transformer.py
@@ -119,8 +119,8 @@ def test_MHA(device, num_heads=12, embed_dim=768, input_seq=512):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_EncoderBlock(device)
     # test_Attention(device, head=16, seq=512, d_k=64)
diff --git a/tests/test_transpose2D.py b/tests/test_transpose2D.py
index 14f16fbb..af5aacf7 100644
--- a/tests/test_transpose2D.py
+++ b/tests/test_transpose2D.py
@@ -46,8 +46,8 @@ def transpose(a, b):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_Transpose2D(device, [64, 156])
     test_Transpose2D_2(device, [16, 64])
diff --git a/tests/test_transpose3D.py b/tests/test_transpose3D.py
index 937948c4..d6c1092d 100644
--- a/tests/test_transpose3D.py
+++ b/tests/test_transpose3D.py
@@ -61,8 +61,8 @@ def transpose(a, b):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_Transpose3D_1(device, [62, 34, 44])
     test_Transpose3D_1(device, [62, 134, 144])
diff --git a/tests/test_vectorops.py b/tests/test_vectorops.py
index 0677b7ae..ed895171 100644
--- a/tests/test_vectorops.py
+++ b/tests/test_vectorops.py
@@ -6,8 +6,8 @@
     import os
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     
     # Target shape
diff --git a/tests/test_view3D_2D.py b/tests/test_view3D_2D.py
index a5a31a85..148fe8fa 100644
--- a/tests/test_view3D_2D.py
+++ b/tests/test_view3D_2D.py
@@ -44,8 +44,8 @@ def view2D_3D(a):
     import sys
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
 
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     test_view3D_2D(device)
     test_view3D_2D(device, [12, 512, 64])
diff --git a/tests/test_vit.py b/tests/test_vit.py
index 6f587127..aeb4f148 100644
--- a/tests/test_vit.py
+++ b/tests/test_vit.py
@@ -202,8 +202,8 @@ def test_encoder_block_with_class_token(
     shape = tuple(map(int, args.shape.strip('()').split(',')))
 
     sys.path.append(os.environ.get('TORCHSIM_DIR', default='/workspace/PyTorchSim'))
-    from Scheduler.scheduler import ExecutionEngine
-    module = ExecutionEngine.setup_device()
+    from Scheduler.scheduler import PyTorchSimRunner
+    module = PyTorchSimRunner.setup_device()
     device = module.custom_device()
     #test_multihead_attention(device)
     #test_encoder_block(device, seq_len=197)