test selecting the correct backend. temp backends while slurm and TE are decoupled (#3848)

* test selecting the correct backend. tem backends while slurm and TE are decoupled

* test selecting the correct backend. tem backends while slurm and TE are decoupled

pytorch_lightning/accelerators/__init__.py

@@ -9,3 +9,5 @@ from pytorch_lightning.accelerators.tpu_backend import TPUBackend
 from pytorch_lightning.accelerators.horovod_backend import HorovodBackend
 from pytorch_lightning.accelerators.ddp_slurm_backend import DDPSLURMBackend
 from pytorch_lightning.accelerators.ddp_torchelastic_backend import DDPTorchElasticBackend
+from pytorch_lightning.accelerators.ddp_cpu_torchelastic_backend import DDPCPUTorchElasticBackend
+from pytorch_lightning.accelerators.ddp_cpu_slurm_backend import DDPCPUSLURMBackend

pytorch_lightning/accelerators/accelerator_connector.py

@@ -158,6 +158,9 @@ class AcceleratorConnector:
         use_ddp_spawn = self.trainer.use_ddp and self.trainer.distributed_backend == "ddp_spawn"
         use_ddp_cpu_spawn = self.trainer.use_ddp and self.trainer.distributed_backend == "ddp_cpu"
 
+        use_ddp_cpu_torch_elastic = use_ddp_cpu_spawn and self._is_using_torchelastic()
+        use_ddp_cpu_slurm = use_ddp_cpu_spawn and self.trainer.is_slurm_managing_tasks
+
         # ddp script mode uses the same flags as TE
         # TODO: decouple from TE
         if os.environ.get('PL_DDP_PID', False):
@@ -167,9 +170,15 @@ class AcceleratorConnector:
         if self.trainer.use_ddp2:
             accelerator_backend = accelerators.DDP2Backend(self.trainer)
 
+        elif use_ddp_cpu_slurm:
+            accelerator_backend = accelerators.DDPCPUSLURMBackend(self.trainer)
+
         elif use_slurm_ddp:
             accelerator_backend = accelerators.DDPSLURMBackend(self.trainer)
 
+        elif use_ddp_cpu_torch_elastic:
+            accelerator_backend = accelerators.DDPCPUTorchElasticBackend(self.trainer)
+
         elif use_torchelastic_ddp:
             accelerator_backend = accelerators.DDPTorchElasticBackend(self.trainer)
 

pytorch_lightning/accelerators/ddp_backend.py

@@ -29,6 +29,7 @@ from pytorch_lightning.utilities.distributed import rank_zero_only
 from pytorch_lightning.utilities import AMPType
 from pytorch_lightning.utilities.seed import seed_everything
 from pytorch_lightning.distributed.dist import LightningDistributed
+from pytorch_lightning.utilities.exceptions import MisconfigurationException
 
 
 try:
@@ -93,6 +94,9 @@ class DDPBackend(Accelerator):
         # when the trainer script was called the device has already been scoped by the time
         # code reaches this point. so, to call the scripts, we need to leave cuda visible devices alone
         # but forward the GPUs selected via environment variables
+        if self.trainer.data_parallel_device_ids is None:
+            raise MisconfigurationException('you selected (distribute_backend = ddp) but did not set Trainer(gpus=?)')
+
         os.environ['PL_TRAINER_GPUS'] = ','.join([str(i) for i in self.trainer.data_parallel_device_ids])
         os.environ['PL_IN_DDP_SUBPROCESS'] = '1'
 

pytorch_lightning/accelerators/ddp_cpu_slurm_backend.py

@@ -0,0 +1,173 @@
+# Copyright The PyTorch Lightning team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License
+import os
+import torch
+import torch.distributed as torch_distrib
+import torch.distributed as dist
+
+from pytorch_lightning.accelerators.base_backend import Accelerator
+from pytorch_lightning import _logger as log
+from pytorch_lightning.utilities import AMPType
+from pytorch_lightning.utilities.distributed import rank_zero_only
+from pytorch_lightning.utilities.seed import seed_everything
+from pytorch_lightning.distributed.dist import LightningDistributed
+
+
+try:
+    from hydra.utils import to_absolute_path, get_original_cwd
+    from hydra.core.hydra_config import HydraConfig
+except ImportError:
+    HYDRA_AVAILABLE = False
+else:
+    HYDRA_AVAILABLE = True
+
+
+# -------------------------------------------
+# !!!!!!!!!!!!!! NOTE !!!!!!!!!!!!!!!!!!!!!!
+# TEMP CLASS WHILE WE DECOUPLE TE FROM DDP
+# !!!!!!!!!!!!!! NOTE !!!!!!!!!!!!!!!!!!!!!!
+# -------------------------------------------
+class DDPCPUSLURMBackend(Accelerator):
+
+    def __init__(self, trainer, cluster_environment=None):
+        super().__init__(trainer, cluster_environment)
+        self.task_idx = None
+        self._has_spawned_children = False
+        self.dist = LightningDistributed()
+
+    def setup(self, model):
+        self.trainer.model = model
+        self.task_idx = int(os.environ['SLURM_LOCALID'])
+
+    def train(self):
+        model = self.trainer.model
+        self.ddp_train(process_idx=self.task_idx, model=model)
+
+    def set_world_ranks(self, process_idx):
+        self.trainer.local_rank = process_idx
+        self.trainer.global_rank = self.trainer.node_rank * self.trainer.num_processes + process_idx
+        self.trainer.world_size = self.trainer.num_nodes * self.trainer.num_processes
+
+    def model_to_device(self, model, process_idx):
+        model.cpu()
+
+    def get_device_ids(self):
+        device_ids = None
+        return device_ids
+
+    def training_step(self, args):
+        if self.trainer.amp_backend == AMPType.NATIVE:
+            with torch.cuda.amp.autocast():
+                output = self.trainer.model(*args)
+        else:
+            output = self.trainer.model(*args)
+        return output
+
+    def validation_step(self, args):
+        output = self.training_step(args)
+        return output
+
+    def test_step(self, args):
+        output = self.training_step(args)
+        return output
+
+    def barrier(self, name: str = None):
+        if torch_distrib.is_initialized():
+            torch_distrib.barrier()
+
+    def early_stopping_should_stop(self, pl_module):
+        stop = torch.tensor(int(self.trainer.should_stop), device=pl_module.device)
+        dist.all_reduce(stop, op=dist.reduce_op.SUM)
+        dist.barrier()
+        should_stop = stop == self.trainer.world_size
+        return should_stop
+
+    def broadcast(self, obj, src=0):
+        return self.dist.broadcast(obj)
+
+    def ddp_train(self, process_idx, model):
+        """
+        Entry point for ddp
+
+        Args:
+            process_idx:
+            mp_queue: multiprocessing queue
+            model:
+
+        Returns:
+
+        """
+        # determine which process we are and world size
+        self.set_world_ranks(process_idx)
+
+        # toggle prog bar
+        if self.trainer.global_rank == 0 and self.trainer.progress_bar_callback is not None:
+            self.trainer.progress_bar_callback.disable()
+
+        # set warning rank
+        rank_zero_only.rank = self.trainer.global_rank
+
+        # set up server using proc 0's ip address
+        # try to init for 20 times at max in case ports are taken
+        # where to store ip_table
+        model.trainer = self.trainer
+        self.init_ddp_connection(
+            self.trainer.global_rank,
+            self.trainer.world_size,
+            self.trainer.is_slurm_managing_tasks
+        )
+
+        # call setup after the ddp process has connected
+        self.trainer.call_setup_hook(model)
+
+        # on world_size=0 let everyone know training is starting
+        if self.trainer.is_global_zero and not torch.distributed.is_initialized():
+            log.info('-' * 100)
+            log.info(f'distributed_backend={self.trainer.distributed_backend} (TORCH_ELASTIC)')
+            log.info(f'All DDP processes registered. Starting ddp with {self.trainer.world_size} processes')
+            log.info('-' * 100)
+
+        # call sync_bn before .cuda(), configure_apex and configure_ddp
+        if self.trainer.sync_batchnorm:
+            model = model.configure_sync_batchnorm(model)
+
+        # move the model to the correct device
+        self.model_to_device(model, process_idx)
+
+        # CHOOSE OPTIMIZER
+        # allow for lr schedulers as well
+        self.setup_optimizers(model)
+
+        # set model properties before going into wrapper
+        self.trainer.model_connector.copy_trainer_model_properties(model)
+
+        # 16-bit
+        model = self.trainer.precision_connector.connect(model)
+
+        # device ids change depending on the DDP setup
+        device_ids = self.get_device_ids()
+
+        # allow user to configure ddp
+        model = model.configure_ddp(model, device_ids)
+
+        # set up training routine
+        self.trainer.train_loop.setup_training(model)
+
+        # train or test
+        results = self.train_or_test()
+
+        # clean up memory
+        torch.cuda.empty_cache()
+
+        return results

pytorch_lightning/accelerators/ddp_cpu_torchelastic_backend.py

@@ -0,0 +1,173 @@
+# Copyright The PyTorch Lightning team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License
+import os
+import torch
+import torch.distributed as torch_distrib
+import torch.distributed as dist
+
+from pytorch_lightning.accelerators.base_backend import Accelerator
+from pytorch_lightning import _logger as log
+from pytorch_lightning.utilities import AMPType
+from pytorch_lightning.utilities.distributed import rank_zero_only
+from pytorch_lightning.utilities.seed import seed_everything
+from pytorch_lightning.distributed.dist import LightningDistributed
+
+
+try:
+    from hydra.utils import to_absolute_path, get_original_cwd
+    from hydra.core.hydra_config import HydraConfig
+except ImportError:
+    HYDRA_AVAILABLE = False
+else:
+    HYDRA_AVAILABLE = True
+
+
+# -------------------------------------------
+# !!!!!!!!!!!!!! NOTE !!!!!!!!!!!!!!!!!!!!!!
+# TEMP CLASS WHILE WE DECOUPLE TE FROM DDP
+# !!!!!!!!!!!!!! NOTE !!!!!!!!!!!!!!!!!!!!!!
+# -------------------------------------------
+class DDPCPUTorchElasticBackend(Accelerator):
+
+    def __init__(self, trainer, cluster_environment=None):
+        super().__init__(trainer, cluster_environment)
+        self.task_idx = None
+        self._has_spawned_children = False
+        self.dist = LightningDistributed()
+
+    def setup(self, model):
+        self.trainer.model = model
+        self.task_idx = int(os.environ['LOCAL_RANK'])
+
+    def train(self):
+        model = self.trainer.model
+        self.ddp_train(process_idx=self.task_idx, model=model)
+
+    def set_world_ranks(self, process_idx):
+        self.trainer.local_rank = process_idx
+        self.trainer.global_rank = self.trainer.node_rank * self.trainer.num_processes + process_idx
+        self.trainer.world_size = self.trainer.num_nodes * self.trainer.num_processes
+
+    def model_to_device(self, model, process_idx):
+        model.cpu()
+
+    def get_device_ids(self):
+        device_ids = None
+        return device_ids
+
+    def training_step(self, args):
+        if self.trainer.amp_backend == AMPType.NATIVE:
+            with torch.cuda.amp.autocast():
+                output = self.trainer.model(*args)
+        else:
+            output = self.trainer.model(*args)
+        return output
+
+    def validation_step(self, args):
+        output = self.training_step(args)
+        return output
+
+    def test_step(self, args):
+        output = self.training_step(args)
+        return output
+
+    def barrier(self, name: str = None):
+        if torch_distrib.is_initialized():
+            torch_distrib.barrier()
+
+    def early_stopping_should_stop(self, pl_module):
+        stop = torch.tensor(int(self.trainer.should_stop), device=pl_module.device)
+        dist.all_reduce(stop, op=dist.reduce_op.SUM)
+        dist.barrier()
+        should_stop = stop == self.trainer.world_size
+        return should_stop
+
+    def broadcast(self, obj, src=0):
+        return self.dist.broadcast(obj)
+
+    def ddp_train(self, process_idx, model):
+        """
+        Entry point for ddp
+
+        Args:
+            process_idx:
+            mp_queue: multiprocessing queue
+            model:
+
+        Returns:
+
+        """
+        # determine which process we are and world size
+        self.set_world_ranks(process_idx)
+
+        # toggle prog bar
+        if self.trainer.global_rank == 0 and self.trainer.progress_bar_callback is not None:
+            self.trainer.progress_bar_callback.disable()
+
+        # set warning rank
+        rank_zero_only.rank = self.trainer.global_rank
+
+        # set up server using proc 0's ip address
+        # try to init for 20 times at max in case ports are taken
+        # where to store ip_table
+        model.trainer = self.trainer
+        self.init_ddp_connection(
+            self.trainer.global_rank,
+            self.trainer.world_size,
+            self.trainer.is_slurm_managing_tasks
+        )
+
+        # call setup after the ddp process has connected
+        self.trainer.call_setup_hook(model)
+
+        # on world_size=0 let everyone know training is starting
+        if self.trainer.is_global_zero and not torch.distributed.is_initialized():
+            log.info('-' * 100)
+            log.info(f'distributed_backend={self.trainer.distributed_backend} (TORCH_ELASTIC)')
+            log.info(f'All DDP processes registered. Starting ddp with {self.trainer.world_size} processes')
+            log.info('-' * 100)
+
+        # call sync_bn before .cuda(), configure_apex and configure_ddp
+        if self.trainer.sync_batchnorm:
+            model = model.configure_sync_batchnorm(model)
+
+        # move the model to the correct device
+        self.model_to_device(model, process_idx)
+
+        # CHOOSE OPTIMIZER
+        # allow for lr schedulers as well
+        self.setup_optimizers(model)
+
+        # set model properties before going into wrapper
+        self.trainer.model_connector.copy_trainer_model_properties(model)
+
+        # 16-bit
+        model = self.trainer.precision_connector.connect(model)
+
+        # device ids change depending on the DDP setup
+        device_ids = self.get_device_ids()
+
+        # allow user to configure ddp
+        model = model.configure_ddp(model, device_ids)
+
+        # set up training routine
+        self.trainer.train_loop.setup_training(model)
+
+        # train or test
+        results = self.train_or_test()
+
+        # clean up memory
+        torch.cuda.empty_cache()
+
+        return results

pytorch_lightning/trainer/connectors/slurm_connector.py

@@ -29,6 +29,10 @@ class SLURMConnector:
                 self.trainer.num_slurm_tasks = int(os.environ['SLURM_NTASKS'])
                 self.trainer.is_slurm_managing_tasks = self.trainer.num_slurm_tasks == self.trainer.num_requested_gpus
 
+                # enable slurm cpu
+                if self.trainer.num_requested_gpus == 0:
+                    self.trainer.is_slurm_managing_tasks = self.trainer.num_slurm_tasks == self.trainer.num_processes
+
                 # in interactive mode we don't manage tasks
                 job_name = os.environ['SLURM_JOB_NAME']
                 if job_name == 'bash':

tests/backends/test_accelerator_connector.py

@@ -0,0 +1,218 @@
+# Copyright The PyTorch Lightning team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License
+
+import pytest
+import os
+from tests.base.boring_model import BoringModel
+from pytorch_lightning.callbacks import Callback
+from pytorch_lightning import accelerators, Trainer
+from unittest import mock
+
+
+def test_accelerator_choice_cpu(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.CPUBackend)
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        callbacks=[CB()]
+    )
+    trainer.fit(model)
+
+
+def test_accelerator_choice_ddp_cpu(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPCPUSpawnBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp_cpu',
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {"CUDA_VISIBLE_DEVICES": "0,1"})
+@mock.patch('torch.cuda.device_count', return_value=2)
+def test_accelerator_choice_ddp(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp',
+        gpus=1,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {"CUDA_VISIBLE_DEVICES": "0,1"})
+@mock.patch('torch.cuda.device_count', return_value=2)
+def test_accelerator_choice_ddp_spawn(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPSpawnBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp_spawn',
+        gpus=1,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {
+    "CUDA_VISIBLE_DEVICES": "0,1",
+    "SLURM_NTASKS": "2",
+    "SLURM_JOB_NAME": "SOME_NAME",
+    "SLURM_NODEID": "0",
+    "SLURM_LOCALID": "0"
+})
+@mock.patch('torch.cuda.device_count', return_value=2)
+def test_accelerator_choice_ddp_slurm(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPSLURMBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp',
+        gpus=2,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {
+    "CUDA_VISIBLE_DEVICES": "0,1",
+    "SLURM_NTASKS": "2",
+    "SLURM_JOB_NAME": "SOME_NAME",
+    "SLURM_NODEID": "0",
+    "LOCAL_RANK": "0",
+    "SLURM_LOCALID": "0"
+})
+@mock.patch('torch.cuda.device_count', return_value=2)
+def test_accelerator_choice_ddp2_slurm(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDP2Backend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp2',
+        gpus=2,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {
+    "CUDA_VISIBLE_DEVICES": "0,1",
+    "WORLD_SIZE": "2",
+    "LOCAL_RANK": "0",
+    "NODE_RANK": "0"
+})
+@mock.patch('torch.cuda.device_count', return_value=2)
+def test_accelerator_choice_ddp_te(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPTorchElasticBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp',
+        gpus=2,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {
+    "WORLD_SIZE": "1",
+    "LOCAL_RANK": "0",
+    "NODE_RANK": "0"
+})
+@mock.patch('torch.cuda.device_count', return_value=0)
+def test_accelerator_choice_ddp_cpu_te(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPCPUTorchElasticBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp_cpu',
+        num_processes=1,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)
+
+
+@mock.patch.dict(os.environ, {
+    "SLURM_NTASKS": "1",
+    "SLURM_JOB_NAME": "SOME_NAME",
+    "SLURM_NODEID": "0",
+    "LOCAL_RANK": "0",
+    "SLURM_LOCALID": "0"
+})
+@mock.patch('torch.cuda.device_count', return_value=0)
+def test_accelerator_choice_ddp_cpu_slurm(tmpdir):
+    class CB(Callback):
+        def on_fit_start(self, trainer, pl_module):
+            assert isinstance(trainer.accelerator_backend, accelerators.DDPCPUSLURMBackend)
+            raise SystemExit()
+
+    model = BoringModel()
+    trainer = Trainer(
+        fast_dev_run=True,
+        distributed_backend='ddp_cpu',
+        num_processes=1,
+        callbacks=[CB()]
+    )
+
+    with pytest.raises(SystemExit):
+        trainer.fit(model)