You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ColossalAI/colossalai/zero/gemini/gemini_hook.py

109 lines
4.2 KiB

from contextlib import contextmanager
from enum import Enum
from functools import partial
from typing import List
import torch
from colossalai.accelerator import get_accelerator
from colossalai.tensor.param_op_hook import ColoParamOpHook
from colossalai.utils import is_ddp_ignored
from colossalai.zero.gemini import TensorState
from colossalai.zero.gemini.gemini_mgr import GeminiManager
class TrainingPhase(Enum):
FORWARD = 0
BACKWARD = 1
class GeminiZeROHook(ColoParamOpHook):
def __init__(self, gemini_manager: GeminiManager) -> None:
super().__init__()
self._gemini_manager = gemini_manager
self._chunk_manager = gemini_manager.chunk_manager
self._training_phase = TrainingPhase.FORWARD
def pre_op(self, params):
# map params to chunks
params = [p for p in params if not is_ddp_ignored(p)]
all_chunks = self._chunk_manager.get_chunks(params)
# wait for prefetched chunks, filter those are not prefetched
chunks_fetch_sync = self._gemini_manager.wait_chunks(all_chunks)
# transfer state
for p in params:
self._chunk_manager.trans_tensor_state(p, TensorState.COMPUTE)
self._gemini_manager.sample_overall_data()
# evit chunks, aware of async fetched
self._gemini_manager.adjust_layout(
all_chunks, record_anyway=self._gemini_manager.placement_policy.max_prefetch > 0
)
# fetch the rest synchronously
for chunk in chunks_fetch_sync:
self._chunk_manager.access_chunk(chunk)
# get possible chunks to prefetch
chunks_fetch_async = self._gemini_manager.placement_policy.get_prefetch_chunks(
is_warmup=self._gemini_manager.is_warmup(),
compute_list=self._gemini_manager.compute_list,
compute_idx=self._gemini_manager.compute_idx,
async_works=self._gemini_manager.async_works,
)
# prefetch
if self._gemini_manager.chunk_manager._prefetch_stream is not None:
# This is when prefetch happens the first time and there is no dist.Work to sync,
# there is possibility that the optimizer haven't finish computation on default stream,
# thus we might prefetch outdated chunks there.
#
# Other than that, self._gemini_manager.wait_chunks will have synced with default stream
# by calling dist.Work.wait() and this line makes no diff.
self._gemini_manager.chunk_manager._prefetch_stream.wait_stream(torch.cuda.current_stream())
with get_accelerator().stream(self._gemini_manager.chunk_manager._prefetch_stream):
for chunk in chunks_fetch_async:
maybe_work = self._chunk_manager.access_chunk(chunk, async_access=True)
if maybe_work is not None:
self._gemini_manager.add_work(chunk, maybe_work)
# record cuda model data of the current OP, including memory for prefetched chunks
self._gemini_manager.record_model_data_volume()
def post_op(self, params):
params = [p for p in params if not is_ddp_ignored(p)]
for p in params:
tensor_state = (
TensorState.HOLD
if self._training_phase == TrainingPhase.FORWARD or not p.requires_grad
else TensorState.HOLD_AFTER_BWD
)
self._chunk_manager.trans_tensor_state(p, tensor_state)
def pre_forward(self, params: List[torch.Tensor]) -> None:
self.pre_op(params)
def post_forward(self, params: List[torch.Tensor]) -> None:
self.post_op(params)
def pre_backward(self, params: List[torch.Tensor]) -> None:
self.pre_op(params)
def post_backward(self, params: List[torch.Tensor]) -> None:
self.post_op(params)
@contextmanager
def switch_training_phase(self, training_phase: TrainingPhase = TrainingPhase.BACKWARD):
old_training_phase = self._training_phase
try:
self._training_phase = training_phase
yield
finally:
self._training_phase = old_training_phase
switch_to_backward = switch_training_phase
switch_to_forward = partial(switch_to_backward, training_phase=TrainingPhase.FORWARD)