Skip to content

Commit

Permalink
[NFC] polish colossalai/gemini/update/chunkv2.py code style (#1565)
Browse files Browse the repository at this point in the history
  • Loading branch information
@zhengzangw @FrankLeeeee
zhengzangw authored and FrankLeeeee committed Sep 8, 2022
1 parent f586887 commit 9823cbf
Showing 1 changed file with 20 additions and 25 deletions.
45 changes: 20 additions & 25 deletions colossalai/gemini/update/chunkv2.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@


class ChunkV2:

def __init__(self,
chunk_size: int,
process_group: ColoProcessGroup,
Expand Down Expand Up @@ -49,9 +50,9 @@ def __init__(self,

self.dtype = dtype
device = init_device or get_current_device()
self.chunk_temp = torch.zeros(chunk_size, dtype=dtype, device=device) # keep all zero
self.chunk_total = None # we force chunk_total located in CUDA
self.cuda_shard = None # using two attributes for the better interpretation
self.chunk_temp = torch.zeros(chunk_size, dtype=dtype, device=device) # keep all zero
self.chunk_total = None # we force chunk_total located in CUDA
self.cuda_shard = None # using two attributes for the better interpretation
self.cpu_shard = None
self.is_gathered = True

Expand All @@ -71,7 +72,7 @@ def __init__(self,
# so their computation patterns are the same as that of the parameters in DDP
self.keep_gathered = keep_gathered
if self.keep_gathered:
pin_memory = False # since this chunk is gathered, it doesn't need to pin
pin_memory = False # since this chunk is gathered, it doesn't need to pin

# if pin_memory is True, we allocate a piece of CPU pin-memory
# for it all the time
Expand Down Expand Up @@ -137,9 +138,9 @@ def append_tensor(self, tensor: torch.Tensor):
if new_utilized_size > self.chunk_size:
raise ChunkFullError

self.chunk_temp[self.utilized_size: new_utilized_size].copy_(tensor.data.flatten())
self.chunk_temp[self.utilized_size:new_utilized_size].copy_(tensor.data.flatten())
assert type(self.chunk_temp) == torch.Tensor, "copy_tensor_to_chunk_slice must use a torch tensor"
tensor.data = self.chunk_temp[self.utilized_size: new_utilized_size].view(tensor.shape)
tensor.data = self.chunk_temp[self.utilized_size:new_utilized_size].view(tensor.shape)

# record all the information about the tensor
self.num_tensors += 1
Expand Down Expand Up @@ -177,11 +178,9 @@ def close_chunk(self, shard_dev: Optional[torch.device] = None):
shard_dev = torch.device('cpu')

if self.pin_memory or shard_dev.type == 'cpu':
self.cpu_shard = torch.empty(self.shard_size,
dtype=self.dtype,
pin_memory=self.pin_memory)
self.cpu_shard = torch.empty(self.shard_size, dtype=self.dtype, pin_memory=self.pin_memory)
self.cpu_shard.copy_(self.cuda_shard)
self.cpu_vis_flag = True # cpu_shard has been visited
self.cpu_vis_flag = True # cpu_shard has been visited

if shard_dev.type == 'cpu':
self.cuda_shard = None
Expand Down Expand Up @@ -260,8 +259,7 @@ def reduce(self):
# we use all-reduce here
dist.all_reduce(self.chunk_total, group=self.torch_pg)
else:
self.cuda_shard = torch.empty(
self.shard_size, dtype=self.dtype, device=get_current_device())
self.cuda_shard = torch.empty(self.shard_size, dtype=self.dtype, device=get_current_device())

input_list = list(torch.chunk(self.chunk_total, chunks=self.pg_size, dim=0))
dist.reduce_scatter(self.cuda_shard, input_list, group=self.torch_pg)
Expand Down Expand Up @@ -330,10 +328,10 @@ def has_inf_or_nan(self) -> bool:
Check if the chunk has inf or nan values in CUDA.
"""
if self.is_gathered:
valid_tensor = self.chunk_total[: self.utilized_size]
valid_tensor = self.chunk_total[:self.utilized_size]
else:
assert self.cuda_shard is not None # only check in CUDA
valid_tensor = self.cuda_shard[: self.valid_end]
assert self.cuda_shard is not None # only check in CUDA
valid_tensor = self.cuda_shard[:self.valid_end]

return torch.isinf(valid_tensor).any().item() | torch.isnan(valid_tensor).any().item()

Expand All @@ -346,8 +344,7 @@ def __gather(self):
self.chunk_total = self.cuda_shard
else:
alloc_storage(self.chunk_total)
gather_list = list(torch.chunk(
input=self.chunk_total, chunks=self.pg_size, dim=0))
gather_list = list(torch.chunk(input=self.chunk_total, chunks=self.pg_size, dim=0))
dist.all_gather(gather_list, self.cuda_shard, self.torch_pg)

self.cuda_shard = None
Expand All @@ -361,11 +358,9 @@ def __scatter(self):
# sanity check
assert self.cuda_shard is None

self.cuda_shard = torch.empty(self.shard_size,
dtype=self.dtype,
device=self.chunk_total.device)
self.cuda_shard = torch.empty(self.shard_size, dtype=self.dtype, device=self.chunk_total.device)

self.cuda_shard.copy_(self.chunk_total[self.shard_begin: self.shard_end])
self.cuda_shard.copy_(self.chunk_total[self.shard_begin:self.shard_end])

free_storage(self.chunk_total)
self.is_gathered = False
Expand Down Expand Up @@ -412,15 +407,15 @@ def __eq__(self, __o: object) -> bool:
def __repr__(self, detailed: bool = False):
output = [
"AgChunk Information:\n",
"\tchunk size: {}, chunk dtype: {}, process group size: {}\n".format(
self.chunk_size, self.dtype, self.pg_size),
"\tchunk size: {}, chunk dtype: {}, process group size: {}\n".format(self.chunk_size, self.dtype,
self.pg_size),
"\t# of tensors: {}, utilized size: {}, utilized percentage: {:.2f}\n".format(
self.num_tensors, self.utilized_size, self.utilized_size / self.chunk_size)
]

def print_tensor(tensor, prefix=''):
output.append("{}shape: {}, dtype: {}, device: {}\n".format(
prefix, tensor.shape, tensor.dtype, tensor.device))
output.append("{}shape: {}, dtype: {}, device: {}\n".format(prefix, tensor.shape, tensor.dtype,
tensor.device))

if self.chunk_temp is not None:
output.append("\tchunk temp:\n")
Expand Down

0 comments on commit 9823cbf

Please sign in to comment.