Summary: The CPU block in `collective_post` was missing pre & post processing. The reduce-scatter implementaion expects use of pre-processing callback to flatten the input tensors, however, the missing invocation meant grabage values were being passed.
Test Plan: Tested the reduce-scatter collective using PARAM
Reviewed By: eastzone
Differential Revision: D41291592
Pull Request resolved: https://github.com/pytorch/pytorch/pull/89030
Approved by: https://github.com/kingchc, https://github.com/kwen2501
Currently all of the distributed errors are thrown from the `TORCH_CHECK` macro which throws a generic `RuntimeError`. This change introduced a new error type `DistBackendError` which derives from `RuntimeError` to signify there was an error with the backend communication library. This allows for better error handling and analysis at higher levels in the stack. Motivation: https://docs.google.com/document/d/1j6VPOkC6znscliFuiDWMuMV1_fH4Abgdq7TCHMcXai4/edit#heading=h.a9rc38misyx8
Changes:
- introduce new error type
- Update `C10D_NCCL_CHECK`
Sample script to demonstrate new error type
```python
# python -m torch.distributed.run --nproc_per_node=2 <script>.py
import torch
import torch.distributed as dist
if __name__ == "__main__":
dist.init_process_group("nccl")
dist.broadcast(torch.tensor([1, 2, 3]).cuda(), 0)
```
Differential Revision: [D40998803](https://our.internmc.facebook.com/intern/diff/D40998803)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/88134
Approved by: https://github.com/rohan-varma
Add sequence number support for UCC, mostly following format of ProcressGroupNCCL.
Pass new test: `test_all_gather_object_subgroup`
Add skips for gather tests: `test_gather_object` and `test_gather_object_subgroup`
cc @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @osalpekar @jiayisuse @H-Huang @kwen2501 @awgu
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85047
Approved by: https://github.com/kwen2501
Currently CUDA UCC barrier is nonblocking with respect to CPU and there is no flag to change it. To make UCC PG barrier behaviour consistent with NCCL PG in this PR barrier has changed to be always blocking.
cc @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @osalpekar @jiayisuse @H-Huang @kwen2501 @awgu
Pull Request resolved: https://github.com/pytorch/pytorch/pull/86961
Approved by: https://github.com/kwen2501
In this PR:
- graph_task stores graph roots on construction so that we can later traverse through the graph
- before the nodes are returned, they needed to be converted from raw_ptr to shared_ptr, and this should be OK because the graph is guaranteed to be alive
Pull Request resolved: https://github.com/pytorch/pytorch/pull/87507
Approved by: https://github.com/albanD
Fixes#87359, which identifies use after free for reverse device maps. This is only in the dynamic RPC feature and not effecting stable RPC code path.
Unfortunately the test `TensorPipeRpcTest.test_dynamic_rpc_existing_rank_can_communicate_with_new_rank_cuda` that is failing is also running into separate issue. I've temporarily disabled some of the test code to investigate the error in asychronously.
Testing plan:
- tested all the dynamic RPC tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/87627
Approved by: https://github.com/rohan-varma
Headers under torch/csrc/distributed may be referened with relative path, e.g., "<c10d/...>". However, relative path cannot be gracefully handled by Meta internal build when the NCCL PG is hipified to support AMD/RCCL because the "hipified" header files are generated in other directories. Moreover, using absolute path for header inclusion is the state-of-the-art in most components in Pytorch. Thus, this patch refactors all header paths in torch/csrc/distributed to be absolute.
See D39835774 for more details about Meta internal complication.
**How to test**: commit 9e5d199 removes -I./torch/csrc/distributed in compile options. Thus use it to verify we don't miss any relative path use of torch/csrc/distributed headers.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85780
Approved by: https://github.com/kumpera, https://github.com/huydhn
Headers under torch/csrc/distributed may be referened with relative path, e.g., "<c10d/...>". However, relative path cannot be gracefully handled by Meta internal build when the NCCL PG is hipified to support AMD/RCCL because the "hipified" header files are generated in other directories. Moreover, using absolute path for header inclusion is the state-of-the-art in most components in Pytorch. Thus, this patch refactors all header paths in torch/csrc/distributed to be absolute.
See D39835774 for more details about Meta internal complication.
**How to test**: commit 9e5d199 removes -I./torch/csrc/distributed in compile options. Thus use it to verify we don't miss any relative path use of torch/csrc/distributed headers.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85780
Approved by: https://github.com/kumpera
### About this PR
* Update the all_reduce op to dispatch to cpu and cuda implementations. Right now they both perform the same logic so this is essentially a no-op.
* Update test to validate that a separate device implementation is not supported.
### About this stack
In the future we will repurpose ProcessGroup to instead contain a list of Backends (ProcessGroupNCCL/Gloo/UCC) and perform dispatching to them based on tensor type. The CPU and CUDA implementations will be updated to have process group select its CPU and CUDA backends respectively.
Differential Revision: [D39506979](https://our.internmc.facebook.com/intern/diff/D39506979)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/83810
Approved by: https://github.com/kwen2501
### About this PR
* Update the broadcast op to dispatch to cpu and cuda implementations. Right now they both perform the same logic so this is essentially a no-op.
* Add test to validate that a separate device implementation is not supported.
### About this stack
In the future we will repurpose ProcessGroup to instead contain a list of Backends (ProcessGroupNCCL/Gloo/UCC) and perform dispatching to them based on tensor type. The CPU and CUDA implementations will be updated to have process group select its CPU and CUDA backends respectively.
Differential Revision: [D38876771](https://our.internmc.facebook.com/intern/diff/D38876771)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/83735
Approved by: https://github.com/kwen2501
Trying to add any deprecation messages we anticipate we need before 1.13 branch cut. Add deprecation message to process group round robin.
## Test
```python
import torch.distributed as dist
if __name__ == "__main__":
pg = dist._round_robin_process_groups(
[
dist.ProcessGroupGloo(dist.TCPStore("localhost", 29500, 1, True), 0, 1)
]
)
```
gives message
```
W0916 16:19:38.367360 68031 ProcessGroupRoundRobin.cpp:12] Warning: ProcessGroupRoundRobin is deprecated and scheduled to be removed after this current release (1.13). Please file an issue on https://github.com/pytorch/pytorch/issues if there are any concerns or issues with this deprecation. (function ProcessGroupRoundRobin)
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85158
Approved by: https://github.com/rohan-varma
