def backward(ctx, grad_output):
if dist.get_backend(group=ctx.group) is dist.Backend.NCCL:
rank = dist.get_rank(group=ctx.group)
world_size = dist.get_world_size(group=ctx.group)
out_size = list(grad_output.size())
if out_size[0] % world_size != 0:
raise RuntimeError(
f'Tensor with dimensions: {out_size} does '
f'not have first dimension divisible by world_size: {world_size}'
)
out_size[0] = out_size[0] // dist.get_world_size(group=ctx.group)
gx = torch.empty(out_size, device=grad_output.device, dtype=grad_output.dtype)
dist._reduce_scatter_base(gx, grad_output, ReduceOp.SUM, ctx.group)
else:
raise RuntimeError("Backend not supported!")
return (None, gx, None)
def reduce_scatter_hook(state: DefaultState, grad: torch.Tensor,
output: torch.Tensor):
r"""
This FSDP communication hook implements ``reduce_scatter`` algorithm for
sharded FSDP strategies and a necessary pre- and post-division of gradients.
Args:
state (DefaultState): State information, configures pre- and post-division factors.
grad (torch.Tensor): An unsharded gradient for the local batch that needs to be
communicated across ranks.
output (torch.Tensor): Stores a single shard of the gradient after ``reduce_scatter``.
"""
# Average grad by pre-division factor.
if state.gradient_predivide_factor > 1:
grad.div_(state.gradient_predivide_factor)
dist._reduce_scatter_base(output, grad, group=state.process_group)
# Average grad's shard by post-division factor.
if state.gradient_postdivide_factor > 1:
output.div_(state.gradient_postdivide_factor)
def reduce_scatter_base(self, collectiveArgs, retFlag=False, pair=False):
retObj = dist._reduce_scatter_base(
output=collectiveArgs.opTensor,
input=collectiveArgs.ipTensor,
group=collectiveArgs.group,
async_op=collectiveArgs.asyncOp,
) # synchronicity is maintained in runColl
if collectiveArgs.asyncOp:
collectiveArgs.waitObj.append(retObj)
if retFlag:
return retObj
def flush(self) -> None:
"""Flush content of the bucket."""
if self.offset == 0:
assert len(self.callbacks) == 0
return
# reduce-scatter bucket
if hasattr(dist, "_reduce_scatter_base"):
dist._reduce_scatter_base( # type: ignore
self.output_shard[:self.offset],
self.data[:, :self.offset].contiguous(),
group=self.group)
else:
dist.reduce_scatter(self.output_shard[:self.offset],
list(self.data[:, :self.offset].unbind(0)),
group=self.group)
# execute post-reduction callbacks
for callback_fn in self.callbacks:
callback_fn()
# reuse input bucket but allocate a fresh output shard
self.data[:, :self.offset].zero_()
self.offset = 0
self.callbacks.clear()
self.output_shard = torch.zeros_like(self.data[0])
def reduce_scatter_async(
self,
input_list: List[Tensor],
group: ProcessGroup,
callback_fn: Optional[Callable] = None,
) -> None:
"""
Reduce-scatter a list of tensors asynchronously, so smaller reductions
can be bucketed together. The given callback (``callback_fn``) will be
called with the reduced result at some later time. Call ``flush()`` to
force all queued ops and callbacks to be executed.
Note that large inputs will be reduced immediately, and this function
may also flush the relevant bucket to make room for ``input_list``.
Args:
input_list (List[Tensor]): list of tensors to reduce-scatter. List
should contain ``group.size()`` tensors and each tensor should
have identical shape, dtype and device.
group (ProcessGroup): process group for reduction
callback_fn (Callable, Optional): callback function to call after
the reduction executes. Function will be called with a single
argument corresponding to the reduced result.
"""
world_size = group.size()
assert (
len(input_list) == world_size
), f"reduce_scatter received {len(input_list)} inputs, expected group.size() ({world_size})"
first_input = input_list[0]
first_input_size = first_input.numel()
bucket_shard_size = self._get_shard_size(first_input.element_size(),
world_size)
if first_input_size > bucket_shard_size:
# TODO: investigate how to avoid using torch.cat (because it seems to be slow for CPU tensors)
# input is too big to fit in the bucket, reduce-scatter directly
output = torch.zeros_like(input_list[0])
if hasattr(dist, "_reduce_scatter_base"):
input_flattened = torch.cat(input_list)
dist._reduce_scatter_base(output, input_flattened,
group=group) # type: ignore
else:
# fallback
dist.reduce_scatter(output, input_list, group=group)
if callback_fn is not None:
callback_fn(output)
return
bucket = self._get_bucket(first_input, group)
if first_input_size > bucket.data.size(1) - bucket.offset:
# not enough space remaining in bucket, flush it now
bucket.flush()
# copy data from input_list into bucket
stacked_input = torch.stack(input_list).view(world_size,
first_input_size)
offset = bucket.offset
bucket.data[:, offset:offset + first_input_size].copy_(stacked_input)
bucket.offset += first_input_size
# callback will be given the reduced result
if callback_fn is not None:
result_view = bucket.output_shard[offset:offset +
first_input_size].view_as(
first_input)
bucket.callbacks.append(functools.partial(callback_fn,
result_view))
def _post_backward_hook(self, param: Parameter, *unused: Any) -> None:
"""
At the start of :func:`_post_backward_hook`, ``param.grad`` contains the
full gradient for the local batch. The reduce-scatter op will replace
``param.grad`` with a single shard of the summed gradient across all
GPUs. This shard will align with the current GPU rank. For example::
before reduce_scatter:
param.grad (GPU #0): [1, 2, 3, 4]
param.grad (GPU #1): [5, 6, 7, 8]
after reduce_scatter:
param.grad (GPU #0): [6, 8] # 1+5, 2+6
param.grad (GPU #1): [10, 12] # 3+7, 4+8
The local GPU's ``optim.step`` is responsible for updating a single
shard of params, also corresponding to the current GPU's rank. This
alignment is created by :func:`_shard_parameters`, which ensures that
the local optimizer only sees the relevant parameter shard.
"""
# First hook callback will see PRE state. If we have multiple params,
# then subsequent hook callbacks will see POST state.
self._assert_state([TrainingState_.BACKWARD_PRE, TrainingState_.BACKWARD_POST])
self.training_state = TrainingState_.BACKWARD_POST
if param.grad is None:
return
if param.grad.requires_grad:
raise RuntimeError(
"FSDP only works with gradients that don't require gradients"
)
self._free_full_params([param])
# Switch to local shard after backward.
self._use_param_local_shard([param])
# Wait for all work in the current stream to finish, then start the
# reductions in post_backward stream.
self._streams["post_backward"].wait_stream(torch.cuda.current_stream())
with torch.cuda.stream(self._streams["post_backward"]):
orig_grad_data = param.grad.data
if self.gradient_predivide_factor > 1:
# Average grad by world_size for consistency with PyTorch DDP.
param.grad.div_(self.gradient_predivide_factor)
if param._is_sharded: # type: ignore[attr-defined]
grad_flatten = torch.flatten(param.grad)
chunks = list(grad_flatten.chunk(self.world_size))
num_pad = self.world_size * chunks[0].numel() - param.grad.numel()
input_flattened = F.pad(grad_flatten, [0, num_pad])
output = torch.zeros_like(chunks[0])
dist._reduce_scatter_base(
output, input_flattened, group=self.process_group
)
if self.gradient_postdivide_factor > 1:
# Average grad by world_size for consistency with PyTorch DDP.
output.div_(self.gradient_postdivide_factor)
param.grad.data = output
else:
# Currently the only way for _is_sharded to be False is if
# world_size == 1. This could be relaxed in the future, e.g,
# no sharding like PyTorch DDP, in which case grads should be
# all-reduced here.
assert (
self.world_size == 1
), "Currently the only way for _is_sharded to be False is \
world_size == 1"
# After _post_backward_hook returns, orig_grad_data will eventually
# go out of scope, at which point it could otherwise be freed for
# further reuse by the main stream while the div/reduce_scatter/copy
# are underway in the post_backward stream. See:
# github.com/NVIDIA/apex/blob/master/apex/parallel/distributed.py
orig_grad_data.record_stream(self._streams["post_backward"])
