Skip to content

Commit

Permalink
ref: unify slurm and TE under backendPlugin 5/n" (#4582)
Browse files Browse the repository at this point in the history 
* ref: unify slurm and TE under backendPlugin 4/n

* ref: unify slurm and TE under backendPlugin 5/n
  • Loading branch information
@williamFalcon @rohitgr7
williamFalcon authored and rohitgr7 committed Nov 21, 2020
1 parent e2c6457 commit 30d3cd6
Show file tree
Hide file tree
Showing 3 changed files with 12 additions and 174 deletions.
14 changes: 9 additions & 5 deletions pytorch_lightning/accelerators/ddp2_accelerator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,22 +11,20 @@
# 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

from pytorch_lightning.utilities.exceptions import MisconfigurationException
from pytorch_lightning.core.lightning import LightningModule
from pytorch_lightning.core.step_result import Result
from pytorch_lightning.distributed.dist import LightningDistributed
from pytorch_lightning import _logger as log
from pytorch_lightning.accelerators.accelerator import Accelerator
from pytorch_lightning.accelerators.accelerator import Accelerator, ReduceOp
from pytorch_lightning.utilities import AMPType
from pytorch_lightning.utilities.distributed import rank_zero_only
from pytorch_lightning.utilities.distributed import rank_zero_only, sync_ddp_if_available
from torch.nn.parallel import DistributedDataParallel
from typing import List, Optional
from typing import List, Optional, Union, Any

try:
from hydra.utils import to_absolute_path, get_original_cwd
Expand Down Expand Up @@ -203,3 +201,9 @@ def configure_sync_batchnorm(self, model: LightningModule) -> LightningModule:
model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model, process_group=None)

return model

def sync_tensor(self,
tensor: Union[torch.Tensor],
group: Optional[Any] = None,
reduce_op: Optional[Union[ReduceOp, str]] = None) -> torch.Tensor:
return sync_ddp_if_available(tensor, group, reduce_op)
170 changes: 2 additions & 168 deletions pytorch_lightning/accelerators/ddp_cpu_hpc_accelerator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,21 +11,7 @@
# 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
from typing import Any, List, Optional, Union

import torch
import torch.distributed as torch_distrib
import torch.distributed as dist
from torch.nn.parallel import DistributedDataParallel

from pytorch_lightning import _logger as log
from pytorch_lightning.accelerators.accelerator import Accelerator, ReduceOp
from pytorch_lightning.core.lightning import LightningModule
from pytorch_lightning.utilities import AMPType
from pytorch_lightning.utilities.distributed import rank_zero_only
from pytorch_lightning.utilities.distributed import sync_ddp_if_available
from pytorch_lightning.distributed.dist import LightningDistributed
from pytorch_lightning.accelerators.ddp_hpc_accelerator import DDPHPCAccelerator


try:
Expand All @@ -37,167 +23,15 @@
HYDRA_AVAILABLE = True


class DDPCPUHPCAccelerator(Accelerator):
class DDPCPUHPCAccelerator(DDPHPCAccelerator):

def __init__(self, trainer, cluster_environment=None, ddp_plugin=None):
super().__init__(trainer, cluster_environment, ddp_plugin)
self.task_idx = None
self._has_spawned_children = False
self.dist = LightningDistributed()
self.nickname = 'ddp_cpu'

def setup(self, model):
self.trainer.model = model
self.task_idx = self.cluster_environment.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: Optional[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:
Dict with evaluation results
"""
# determine which process we are and world size
self.set_world_ranks(process_idx)

# toggle prog bar
if (self.trainer.node_rank != 0 or process_idx != 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 = self.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 = self.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

def configure_ddp(
self, model: LightningModule, device_ids: List[int]
) -> DistributedDataParallel:
model = self.ddp_plugin.configure_ddp(model, device_ids)
return model

def configure_sync_batchnorm(self, model: LightningModule) -> LightningModule:
"""
Add global batchnorm for a model spread across multiple GPUs and nodes.
Override to synchronize batchnorm between specific process groups instead
of the whole world or use a different sync_bn like `apex`'s version.
Args:
model: pointer to current :class:`LightningModule`.
Return:
LightningModule with batchnorm layers synchronized between process groups
"""
model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model, process_group=None)

return model

def sync_tensor(self,
tensor: Union[torch.Tensor],
group: Optional[Any] = None,
reduce_op: Optional[Union[ReduceOp, str]] = None) -> torch.Tensor:
return sync_ddp_if_available(tensor, group, reduce_op)
2 changes: 1 addition & 1 deletion pytorch_lightning/accelerators/ddp_hpc_accelerator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -136,7 +136,7 @@ def ddp_train(self, process_idx, 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} (on SLURM)')
log.info(f'distributed_backend={self.trainer.distributed_backend}')
log.info(f'All DDP processes registered. Starting ddp with {self.trainer.world_size} processes')
log.info('-' * 100)

Expand Down

0 comments on commit 30d3cd6

Please sign in to comment.