Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/88330
### Implementation
Move backend-specific (NCCL, Gloo, etc) collective implementations to corresponding `Backend` class. Update ProcessGroup to support multiple backends and use dispatcher to calls backends based on tensor device type.
### Changes
#### c++ changes (ProcessGroup files, `Ops.cpp`, `init.cpp`)
- Update pybind definitions for new process group base class and new backend class
- Update pybinded backend class with collective definitions to keep BC with Python PG instances (e.g. `dist.ProcessGroupGloo`, `dist.ProcessGroupNCCL`) which are used in tests
- Switch `ProcessGroupGloo`, `ProcessGroupNCCL`, `ProcessGroupMPI`, `ProcessGroupUCC` to derive from the `Backend` class.
- Update CPU/CUDA `Ops.cpp` and `OpsImpl.cpp` to perform this dispatching by querying the backend using the device type
- Update internal dispatched implementation of `barrier` to use a tensor which allows operation to be dispatched.
- Update `allgather` collective to use `TensorList`. For some reason it was using the default implementation of `allgather` rather than dispatching it correctly. I still don't understand why and had originally filed an issue in 85122.
#### python changes (`distributed_c10d.py`, test files)
- Add BackendConfig class to specify the default configurations of backends and `get_backend_config()` API
- `get_backend()` deprecation warning
- `init_process_group` how returns a generic `ProcessGroup` object, it contains a list of backends (the ones stated above) which it will dispatch operations to.
- `new_group` updated to return the same as above
- Update `test_c10d_gloo.py`, Update `DistributedDataParallelTest` to use `init_process_group`, Update `ReducerTest`, update `test_broadcast_coalesced_gloo` to move from PG instance and gloo options
- Update `test_c10d_nccl.py`, Update `DistributedDataParallelTest` to use `init_process_group`
- Specific tests updated: `test_Backend_enum_class`
### Changes missing
- lazy initialization of backends
- support parsing of BackendConfig
### open questions
- Pure Python PG extensions (https://github.com/pytorch/pytorch/pull/66338)
# Example
This is a basic script (using 2 backends within a process group)
```python
# python -m torch.distributed.run --nnodes=1 --nproc_per_node=2 basic_scenario.py
import torch.distributed as dist
import torch
import os
if __name__ == "__main__":
rank = os.environ.get("RANK")
# initialize with both gloo and nccl
dist.init_process_group()
# with gloo
dist.all_reduce(torch.tensor([1.0]))
print(f"Rank {rank} finished")
# with nccl
dist.all_reduce(torch.tensor([1.0], device=f"cuda:{rank}"))
```
Test Plan: Imported from OSS
Differential Revision: D42069829
Pulled By: H-Huang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/90997
Approved by: https://github.com/awgu, https://github.com/fduwjj
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/67050
This PR moves init_multi_gpu_helper to common_distributed so that it could be shared by different distributed tests.
ghstack-source-id: 141370119
Test Plan: wait for ci.
Reviewed By: mrshenli
Differential Revision: D31842644
fbshipit-source-id: c7bad25d6cef9bdce7ad1fb6c60c1cad4b765702
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63768
passed number of buckets to GradBucket constructor, to check if index is equal to num_buckets - 1 in the .is_last() function.
Test Plan:
buck test mode/dev-nosan //caffe2/test/distributed/algorithms/ddp_comm_hooks:test_ddp_hooks
test output: https://www.internalfb.com/intern/testinfra/testconsole/testrun/8162774375985873/
Reviewed By: SciPioneer, mrshenli
Differential Revision: D30455913
fbshipit-source-id: 8c67ca69cbf191d6e189e09248407eb167bb24b6
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63154
The collective quantization API now supports alltoall, alltoall_single, and allscatter. The test is also included.
ghstack-source-id: 136856877
Test Plan: buck test mode/dev-nosan //caffe2/test/distributed/algorithms/quantization:DistQuantizationTests_nccl -- test_all_to_all_single_bfp16
Reviewed By: wanchaol
Differential Revision: D30255251
fbshipit-source-id: 856f4fa12de104689a03a0c8dc9e3ecfd41cad29
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63059
Supporting BFP16 quantization method to OSS. Currently only support CPU
ghstack-source-id: 136639528
Test Plan: Imported from OSS
Reviewed By: wanchaol
Differential Revision: D30194538
fbshipit-source-id: ac248567ad8028457c2a91b77ef2ce81709fce53
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62051
The goal here is to enable opt-asan for "spawn" based unit tests since
this works for "spawn" unlike "dev-asan". As a result, we can run ASAN for
"spawn" unit tests as well.
This means we can completely remove fork unit tests from the code base since
the only purpose for these tests was to run ASAN.
ghstack-source-id: 135523770
Test Plan: waitforbuildbot
Reviewed By: SciPioneer
Differential Revision: D29854514
fbshipit-source-id: 02a5bfcfae2afc21badecff77082c7a6ad83636b
Summary:
**Overview:**
This removes the preceding `_` from `_Join`, `_Joinable`, and `_JoinHook` in preparation for adding the generic join context manager tutorial (see [here](https://github.com/pytorch/tutorials/pull/1610)). This also adds a docs page, which can be linked from the tutorial. [Here](https://github.com/pytorch/pytorch/files/6919475/render.pdf) is a render of the docs page.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62605
Test Plan:
`DistributedDataParallel.join()`:
```
touch /tmp/barrier && TEMP_DIR="/tmp" BACKEND="nccl" WORLD_SIZE="2" gpurun python test/distributed/test_distributed_fork.py -- TestDistBackendWithFork.test_ddp_uneven_inputs TestDistBackendWithFork.test_ddp_uneven_inputs_stop_iteration_sync_bn TestDistBackendWithFork.test_ddp_grad_div_uneven_inputs TestDistBackendWithFork.test_ddp_uneven_input_join_disable TestDistBackendWithFork.test_ddp_uneven_input_exception
```
`ZeroRedundancyOptimizer`:
```
gpurun4 python test/distributed/optim/test_zero_redundancy_optimizer.py
```
NOTE: DDP overlap tests are failing due to a landing race. See https://github.com/pytorch/pytorch/pull/62592. Once the fix is landed, I will rebase, and tests should be passing.
`Join`:
```
gpurun4 python test/distributed/algorithms/test_join.py
```
Reviewed By: mrshenli
Differential Revision: D30055544
Pulled By: andwgu
fbshipit-source-id: a5ce1f1d9f1904de3bdd4edd0b31b0a612d87026
Summary:
**Overview:**
The existing `test_ddp_hooks.py` test file uses a prefix `c10d`, which is not defined in the file, meaning the test errors if left as is. This renames each `c10d` prefix to `dist`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/61864
Test Plan:
All four tests pass when run:
```
gpurun python test/distributed/algorithms/ddp_comm_hooks/test_ddp_hooks.py
```
Reviewed By: ejguan
Differential Revision: D29783860
Pulled By: andwgu
fbshipit-source-id: 16bdd2dfcb76192964246148f14851a74f8907c8
Summary:
Currently there is some code that intends to skip distributed tests if
the distributed module is not built. However, they are missing in some
test files; and in some other test files they are checked after
distributed module is imported, which leads to failure. This is
generating a lot of headaches when testing minimal builds locally.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52945
Reviewed By: anjali411
Differential Revision: D26848241
Pulled By: ezyang
fbshipit-source-id: 983a848844add40869a86f3c9413503a3659b115
Summary:
Re-enabling these test cases for ROCm because they are passing.
jeffdaily
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52403
Reviewed By: jbschlosser, SciPioneer
Differential Revision: D26516727
Pulled By: malfet
fbshipit-source-id: 6c70805eda39b0aadfbeb30a527af3906d2da867
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43310
In this diff, we prepared some example DDP communication hooks [#40848](https://github.com/pytorch/pytorch/pull/40848):
1\. `allreduce_hook`: This DDP communication hook just calls ``allreduce`` using ``GradBucket`` tensors. Once gradient tensors are aggregated across all workers, its ``then`` callback takes the mean and returns the result. If user registers this hook DDP results is expected to be same as the case where no hook was registered. Hence, this won't change behavior of DDP and user can use this as a reference or modify this hook to log useful information or any other purposes while unaffecting DDP behavior.
2\. `allgather_then_aggregate_hook` Similar to ``allreduce_hook``, this hook first gathers ``GradBucket`` tensors and its ``then`` callback aggregates the gathered gradient tensors and takes mean. Instead of ``allreduce`` this hook uses ``allgather``. Note that with W workers, both the computation and communication time scale as O(W) for allgather compared to O(logW) for allreduce. Therefore, this hook is expected to be much slower than ``allreduce_hook`` although both essentially do the same thing with the gradients.
3\. `fp16_compress_hook` This DDP communication hook implements a simple gradient compression approach that converts ``GradBucket`` tensors whose type is assumed to be ``torch.float32`` to half-precision floating point format (``torch.float16``). It allreduces those ``float16`` gradient tensors. Once compressed gradient tensors are allreduced, its then callback called ``decompress`` converts the aggregated result back to ``float32`` and takes the mean.
4\. `quantization_pertensor_hook` does quantization per tensor and uses the idea in https://pytorch.org/docs/master/generated/torch.quantize_per_tensor.html. Note that we separately send scale and zero_point (two floats per rank) before quantized tensors.
5\. `quantization_perchannel_hook` does quantization per channel similar to https://pytorch.org/docs/master/generated/torch.quantize_per_channel.html. The main motivation is that after the initial QSGD study diff, we realized that for considerably large gradient tensors such as a tensor that contains 6 million floats quantizing dividing it into smaller channels (512 float chunks) and quantizing independently may significantly increase the resolution and result with lower error.
ghstack-source-id: 110923269
Test Plan:
python torch/distributed/algorithms/ddp_comm_hooks/test_ddp_hooks.py
Couldn't download test skip set, leaving all tests enabled...
.....
----------------------------------------------------------------------
Ran 4 tests in 26.724s
OK
Internal testing:
```
buck run mode/dev-nosan //caffe2/test/distributed/algorithms/ddp_comm_hooks:test_ddp_hooks
```
Reviewed By: malfet
Differential Revision: D22937999
fbshipit-source-id: 274452e7932414570999cb978ae77a97eb3fb0ec
