zdou0830's picture
desco
749745d
"""
This file contains primitives for multi-gpu communication.
This is useful when doing distributed training.
"""
import pickle
import time
import functools
import logging
import torch
import torch.distributed as dist
import numpy as np
def get_world_size():
if not dist.is_available():
return 1
if not dist.is_initialized():
return 1
return dist.get_world_size()
def get_rank():
if not dist.is_available():
return 0
if not dist.is_initialized():
return 0
return dist.get_rank()
def is_main_process():
return get_rank() == 0
def synchronize():
"""
Helper function to synchronize (barrier) among all processes when
using distributed training
"""
if not dist.is_available():
return
if not dist.is_initialized():
return
world_size = dist.get_world_size()
if world_size == 1:
return
dist.barrier()
def all_gather(data):
"""
Run all_gather on arbitrary picklable data (not necessarily tensors)
Args:
data: any picklable object
Returns:
list[data]: list of data gathered from each rank
"""
world_size = get_world_size()
if world_size == 1:
return [data]
# serialized to a Tensor
buffer = pickle.dumps(data)
storage = torch.ByteStorage.from_buffer(buffer)
tensor = torch.ByteTensor(storage).to("cuda")
# obtain Tensor size of each rank
local_size = torch.LongTensor([tensor.numel()]).to("cuda")
size_list = [torch.LongTensor([0]).to("cuda") for _ in range(world_size)]
dist.all_gather(size_list, local_size)
size_list = [int(size.item()) for size in size_list]
max_size = max(size_list)
# receiving Tensor from all ranks
# we pad the tensor because torch all_gather does not support
# gathering tensors of different shapes
tensor_list = []
for _ in size_list:
tensor_list.append(torch.ByteTensor(size=(max_size,)).to("cuda"))
if local_size != max_size:
padding = torch.ByteTensor(size=(max_size - local_size,)).to("cuda")
tensor = torch.cat((tensor, padding), dim=0)
dist.all_gather(tensor_list, tensor)
data_list = []
for size, tensor in zip(size_list, tensor_list):
buffer = tensor.cpu().numpy().tobytes()[:size]
data_list.append(pickle.loads(buffer))
return data_list
def reduce_dict(input_dict, average=True):
"""
Args:
input_dict (dict): all the values will be reduced
average (bool): whether to do average or sum
Reduce the values in the dictionary from all processes so that process with rank
0 has the averaged results. Returns a dict with the same fields as
input_dict, after reduction.
"""
world_size = get_world_size()
if world_size < 2:
return input_dict
with torch.no_grad():
names = []
values = []
# sort the keys so that they are consistent across processes
for k in sorted(input_dict.keys()):
names.append(k)
values.append(input_dict[k])
values = torch.stack(values, dim=0)
dist.reduce(values, dst=0)
if dist.get_rank() == 0 and average:
# only main process gets accumulated, so only divide by
# world_size in this case
values /= world_size
reduced_dict = {k: v for k, v in zip(names, values)}
return reduced_dict
def broadcast_data(data):
if not torch.distributed.is_initialized():
return data
rank = dist.get_rank()
if rank == 0:
data_tensor = torch.tensor(data + [0], device="cuda")
else:
data_tensor = torch.tensor(data + [1], device="cuda")
torch.distributed.broadcast(data_tensor, 0)
while data_tensor.cpu().numpy()[-1] == 1:
time.sleep(1)
return data_tensor.cpu().numpy().tolist()[:-1]
def reduce_sum(tensor):
if get_world_size() <= 1:
return tensor
tensor = tensor.clone()
dist.all_reduce(tensor, op=dist.ReduceOp.SUM)
return tensor
def shared_random_seed():
"""
Returns:
int: a random number that is the same across all workers.
If workers need a shared RNG, they can use this shared seed to
create one.
All workers must call this function, otherwise it will deadlock.
"""
ints = np.random.randint(2**31)
all_ints = all_gather(ints)
return all_ints[0]