Skip to content

Commit

Permalink
feat(optimizer/hybrid_zero_optim.py): resolve conflicts
Browse files Browse the repository at this point in the history
  • Loading branch information
@huangting4201
huangting4201 committed Oct 20, 2023
2 parents d91a5d9 + 95488d8 commit 3c69254
Show file tree
Hide file tree
Showing 2 changed files with 34 additions and 78 deletions.
1 change: 0 additions & 1 deletion internlm/core/scheduler/no_pipeline_scheduler.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -194,7 +194,6 @@ def forward_backward_step(
_output, _loss, _moe_loss = self._train_one_batch(
_data, _label, engine, forward_only, return_loss, self._grad_accum_size
)
engine.optimizer.reset_reduce_bucket()

if return_loss:
loss += _loss
Expand Down
111 changes: 34 additions & 77 deletions internlm/solver/optimizer/hybrid_zero_optim.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@

import math
from functools import partial
from typing import List, Optional

import torch
import torch.distributed as dist
Expand Down Expand Up @@ -105,8 +106,8 @@ def __init__(
# it will not manage the tensors used by mixed precision training
self._param_store = ParameterStore(ParallelMode.ZERO1)
self._grad_store = GradientStore(ParallelMode.DATA)
self._bucket_store = []
self._bucket_store_2 = []
self._bucket_store: List[BucketStore] = []
self._accum_grad_buckets: List[BucketStore] = []
self._bucket_in_progress = []

# fp16 and fp32 params for mixed precision training
Expand Down Expand Up @@ -175,7 +176,7 @@ def __init__(
# TODO _broadcast_parallel_mode is not only used in broadcast, maybe can change its name
self._broadcast_parallel_mode.append(zero_mode)
self._bucket_store.append(BucketStore(group_id, param_group["dp_mode"]))
self._bucket_store_2.append(BucketStore(group_id, param_group["dp_mode"]))
self._accum_grad_buckets.append(BucketStore(group_id, param_group["dp_mode"]))

# assign parameters to ranks the params in the list are sorted
params_per_rank, no_params_ranks = self._partition_param_list(group_id, param_group)
Expand Down Expand Up @@ -323,7 +324,7 @@ def _define_and_attach(param, reduce_rank=None):
)

reduce_scatter_checker = partial(
self._wait_reduce_scatter_and_accumulate_grad,
self._wait_reduce_scatter_and_accumulate_grads,
param=param,
reduce_rank=reduce_rank,
)
Expand All @@ -332,7 +333,7 @@ def _define_and_attach(param, reduce_rank=None):
# NOT IMPORTANT BUT GOOD TO KNOW:
# args here is not grad, but allow_unreacable and accumulate_grad
def reduce_grad_hook(*args): # pylint: disable=W0613
if gpc.config.fstp_handler is not None:
if self._fstp_handler is not None:
reduce_scatter_checker()

if self.skip_grad_reduce is False:
Expand All @@ -356,84 +357,36 @@ def belongs_to_current_rank(self, param) -> bool:
group_id = getattr(param, "group_id")
return tensor_rank == gpc.get_local_rank(self._broadcast_parallel_mode[group_id])

def reset_reduce_bucket(self) -> None:
for bucket in self._bucket_store_2:
for rank, params in bucket._params.items():
for _param in params:
if not hasattr(_param, "_fstp_reduce_scatter_str"):
continue

key = getattr(_param, "_fstp_reduce_scatter_str")
comm_handle, _grad = self._fstp_handler.reduce_scatter_handlers[key]
comm_handle.wait()
_param.grad.add_(_grad)
# self._fstp_handler.reduce_scatter_handlers[key] = None
# del _grad
release_reduce_scatter_memory_pool(size=tuple(_grad.size()), index=_grad.index)
del self._fstp_handler.reduce_scatter_handlers[key]
self._fstp_handler.reduce_scatter_handlers[key] = None
assert key in self._fstp_handler.reduce_scatter_handlers
# if not hasattr(_param, "_fstp_all_reduce_str"):
# continue

# key = getattr(_param, "_fstp_all_reduce_str")
# comm_handle, _grad = self._fstp_handler.all_reduce_handlers[key]
# comm_handle.wait()
# with torch.no_grad():
# _grad = split_forward_gather_backward(_grad, ParallelMode.TENSOR, dim=0)
# _param.grad.add_(_grad)
# # self._fstp_handler.reduce_scatter_handlers[key] = None
# del _grad
# del self._fstp_handler.all_reduce_handlers[key]
# self._fstp_handler.all_reduce_handlers[key] = None
# assert key in self._fstp_handler.all_reduce_handlers

bucket.reset_by_rank(rank)

def _wait_reduce_scatter_and_accumulate_grad(self, param, reduce_rank=None):
def _accum_grads_store_in_bucket(self, bucket: BucketStore, reduce_rank: Optional[int] = None) -> None:
for _param in bucket.get_param(reduce_rank):
if not hasattr(_param, "_fstp_reduce_scatter_str"):
continue

# wait and accumulate gardient.
_key = getattr(_param, "_fstp_reduce_scatter_str")
_comm_handle, _grad = self._fstp_handler.reduce_scatter_handlers[_key]
_comm_handle.wait()
_param.grad.add_(_grad)

# release cuda memory.
release_reduce_scatter_memory_pool(size=tuple(_grad.size()), index=_grad.index)
self._fstp_handler.reduce_scatter_handlers[_key] = None

bucket.reset_by_rank(reduce_rank)

def _wait_reduce_scatter_and_accumulate_grads(self, param, reduce_rank: Optional[int] = None):
param_size = param.numel()

group_id = getattr(param, "group_id")
current_bucket = self._accum_grad_buckets[group_id]

# check if the bucket is full
# if full, will reduce the grads already in the bucket
# after reduction, the bucket will be empty
group_id = getattr(param, "group_id")
current_bucket = self._bucket_store_2[group_id]

if current_bucket.num_elements_in_bucket(reduce_rank) >= 512 * 1024 * 1024:
# wait reduce scatter communication
params = current_bucket.get_param(reduce_rank)
for _param in params:
if not hasattr(_param, "_fstp_reduce_scatter_str"):
continue

key = getattr(_param, "_fstp_reduce_scatter_str")
comm_handle, _grad = self._fstp_handler.reduce_scatter_handlers[key]
comm_handle.wait()
_param.grad.add_(_grad)
# self._fstp_handler.reduce_scatter_handlers[key] = None
# del _grad
release_reduce_scatter_memory_pool(size=tuple(_grad.size()), index=_grad.index)
del self._fstp_handler.reduce_scatter_handlers[key]
self._fstp_handler.reduce_scatter_handlers[key] = None
assert key in self._fstp_handler.reduce_scatter_handlers

# if not hasattr(_param, "_fstp_all_reduce_str"):
# continue

# key = getattr(_param, "_fstp_all_reduce_str")
# comm_handle, _grad = self._fstp_handler.all_reduce_handlers[key]
# comm_handle.wait()
# with torch.no_grad():
# _grad = split_forward_gather_backward(_grad, ParallelMode.TENSOR, dim=0)
# _param.grad.add_(_grad)
# # self._fstp_handler.reduce_scatter_handlers[key] = None
# del _grad
# del self._fstp_handler.all_reduce_handlers[key]
# self._fstp_handler.all_reduce_handlers[key] = None
# assert key in self._fstp_handler.all_reduce_handlers

current_bucket.reset_by_rank(reduce_rank)
if current_bucket.num_elements_in_bucket(reduce_rank) >= self._reduce_bucket_size:
self._accum_grads_store_in_bucket(current_bucket, reduce_rank)

# otherwise, add the parameter into bucket.
current_bucket.add_num_elements_in_bucket(param_size, reduce_rank)
current_bucket.add_param(param, reduce_rank)

Expand Down Expand Up @@ -661,6 +614,10 @@ def step(self, closure=None):
for group_id in range(self.num_param_groups):
self._reduce_grads_stored_in_bucket(self._bucket_store[group_id], reduce_rank=None, last_bucket=True)

# we need to accumulate gradients left in the accumulate gardient bucket
for group_id in range(self.num_param_groups):
self._accum_grads_store_in_bucket(self._accum_grad_buckets[group_id], reduce_rank=None)

# compute norm for gradients in the before bucket
groups_norms = []
for group_id in range(self.num_param_groups):
Expand Down

0 comments on commit 3c69254

Please sign in to comment.