[devops] add large-scale distributed test marker (#4452)

* [test] remove cpu marker

* [test] remove gpu marker

* [test] update pytest markers

* [ci] update unit test ci

tests/test_config/test_load_config.py

@@ -8,7 +8,6 @@ import pytest
 from colossalai.context.config import Config
 
 
-@pytest.mark.cpu
 def test_load_config():
     filename = Path(__file__).parent.joinpath('sample_config.py')
     config = Config.from_file(filename)

tests/test_context/test_hybrid_parallel.py

@@ -143,7 +143,6 @@ def run_dist(rank, world_size, port, backend, port_list, host):
         reset_seeds()
 
 
-@pytest.mark.cpu
 @rerun_if_address_is_in_use()
 def test_context():
     """

tests/test_data/test_cifar10_dataset.py

@@ -5,11 +5,10 @@ import os
 from pathlib import Path
 
 import pytest
-from torchvision import transforms, datasets
 from torch.utils.data import DataLoader
+from torchvision import datasets, transforms
 
 
-@pytest.mark.cpu
 def test_cifar10_dataset():
     # build transform
     transform_pipeline = [transforms.ToTensor()]

tests/test_data/test_data_parallel_sampler.py

@@ -53,7 +53,6 @@ def run_data_sampler(rank, world_size, port):
     torch.cuda.empty_cache()
 
 
-@pytest.mark.cpu
 @rerun_if_address_is_in_use()
 def test_data_sampler():
     spawn(run_data_sampler, 4)

tests/test_data/test_deterministic_dataloader.py

@@ -64,7 +64,6 @@ def run_data_sampler(rank, world_size, port):
     torch.cuda.empty_cache()
 
 
-@pytest.mark.cpu
 @rerun_if_address_is_in_use()
 def test_data_sampler():
     spawn(run_data_sampler, 4)

tests/test_utils/test_activation_checkpointing.py

@@ -40,7 +40,6 @@ def forward_inplace(x, weight):
     return out
 
 
-@pytest.mark.gpu
 @clear_cache_before_run()
 @parameterize("use_reentrant", [True, False])
 @parameterize("cpu_offload", [True, False])