This PR mimics what we have done to trace ProcessGroup. This allows use to use FakeProcessGroup with torch.compile. FakeProcessGroup allows us to use world_size > 1 without creating multiple processes thus enabling the usage of PDB to debug bucketing DDP allreduce in the Inductor. We can theoretically use GLOO with world_size==1 to achieve the same goal. However, the `wait()` seems to be optimized away when the world_size is 1.
Differential Revision: [D51136463](https://our.internmc.facebook.com/intern/diff/D51136463/)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113314
Approved by: https://github.com/wanchaol
This PR adds initial dynamo support for DTensor, in particular, it:
- allows DTensor be passed into a compiled function, and allow fakify
DTensor during dynamo tracing by turning the inner local tensor to meta
tensor.
- We use `allow_in_graph` to include `DTensor` and `DTensor.from_local` to be represented as `TorchVariable`
- The dtensor created becomes a normal `TensorVariable` and it would insert any tensor operations to the output graph just like torch.Tensor
- note that dtensor have a new instance method `redistribute` compare to plain tensor, and we currently special handle it in `TensorVariable`
`from_local` and `redistribute` both accepts some non-trival metadata as arguments (i.e. DeviceMesh, Placement) which fx.Graph does not support. In order to let these two APIs appear in the dynamo captured graph, we encoded the metadata into a new_function (like `functools.partial`) and the new function only accepts prim args (i.e. tensor), then we put `call_function` with this new_function to the graph. This is suggested by @ezyang. The underlying rationale here is that the metadata will not change across the graph invocations so it's safe to encode them.
Captured graph:
```
def forward(self, L_x_ : torch.Tensor):
l_x_ = L_x_
# File: /scratch/wanchaol/work/pytorch/test/distributed/_tensor/test_dtensor.py:685, code: dt = DTensor.from_local(x, mesh, [Shard(0)], run_check=False)
prim_from_local = torch__dynamo_variables_torch_prim_from_local(l_x_, run_check = False); l_x_ = None
# File: /scratch/wanchaol/work/pytorch/test/distributed/_tensor/test_dtensor.py:686, code: return dt.redistribute(mesh, [Replicate()]).to_local() + 2
prim_redistribute = torch__dynamo_variables_tensor_prim_redistribute(prim_from_local); prim_from_local = None
to_local = prim_redistribute.to_local(); prim_redistribute = None
add = to_local + 2; to_local = None
return (add,)
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/103146
Approved by: https://github.com/voznesenskym
