These tests are flaky on ROCm and have been skipped via Github issues, but the bot keeps closing the issues after not observing the failures for these tests in the rerun_disabled_tests runs (not sure why they don't fail there), and we have to keep reopening them.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/155724
Approved by: https://github.com/jeffdaily
Co-authored-by: Jithun Nair <37884920+jithunnair-amd@users.noreply.github.com>
Summary: The error caused by the world size not being divisible by `group_size` is a common issue encountered by end-users when utilizing applications built on top of `new_subgroups()`. However, these applications may employ different variable names, such as `num_trainers_per_group`, which can make the current error messages less effective despite being correct. To address this, we have improved the error messages to display the actual numbers involved, thereby enhancing their clarity and usefulness.
Test Plan: contbuild & OSS CI
Differential Revision: D75226925
Pull Request resolved: https://github.com/pytorch/pytorch/pull/154124
Approved by: https://github.com/wz337
Differential Revision: D72437251
Enable to rebuild bucket order when find_unused_parameters=true.
It should be always better than not rebuilding bucket order when find_unused_parameters=True:
1. for cases where bucket order in the first iteration is the same as the parameter order, rebuilding bucket order will not change anything
2. for cases where bucket order in the first iteration is not the same as the parameter order, there could be two cases:
a. bucket order will not change after 1st iteration even the graph is dynamic and there is unused parameter, in this case, rebuilding bucket order will have performance gain
b. bucket order change after 1st iteration due to dynamic graph, in this case, both parameter order and bucket order in 1st iteration are not ideal, so rebuilding bucket order or not does not matter
it can help case 2.a if enabling to rebuild bucket order when find_unused_parameters=true. meanwhile it will not hurt other cases in 1 and 2.b.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153404
Approved by: https://github.com/rohan-varma, https://github.com/fegin
Summary: The bug, introduced in https://github.com/pytorch/pytorch/pull/152765, was caused by passing the `group` parameter to the `get_rank()` function, which caused the function to return the rank of the entire group instead of the rank of the current process. The fix involves removing the `group` parameter from the `get_rank()` function call.
Test Plan: contbuild & OSS CI
Differential Revision: D74964213
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153798
Approved by: https://github.com/Skylion007
MOTIVATION
This PR includes a minor change to check for TEST_HPU flag as well before falling back to CPU. Without this flag, some tests were falling back to CPU causing them to fail.
Please refer to this RFC as well: https://github.com/pytorch/rfcs/pull/66
CHANGES
add TEST_HPU flag to some of the conditions checking the environment
use DEVICE_COUNT variable instead of torch.accelerator.device_count() API since the later is not supported on out-of-tree devices like Intel Gaudi.
@ankurneog , @EikanWang , @cyyever , @guangyey
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153461
Approved by: https://github.com/EikanWang, https://github.com/cyyever, https://github.com/albanD
Summary: add one option to allow skipping all reduce unused parameters, this could help improve training throughput significantly when the number of unused parameters is large in the model.
Test Plan: unit tests, CI
Differential Revision: D72282069
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151503
Approved by: https://github.com/mrshenli
Relanding #148590 due to merge conflict.
This PR has multiple changes to `ProcessGroupNCCL` (which unfortunately are related):
1. When async_op=False, we directly launch the collective on "current" stream, instead of a trampoline stream and join back.
- Resolves#147729
- Resolves#146881
- Also saves two event syncs (which have overhead in case of HIP) and one pybind when we call `work.wait()` in distributed_c10d.py on behalf of user.
2. Entirely remove `record_stream` and use CPU-side stashing for managing tensor lifetime against recycling.
- Resolves#147168
3. Remove tensor life management when async_op=False; only use it when async_op=True.
4. To guard against user not calling `work.wait()`, we ask watchdog to unstash tensors after detecting completion of collectives, to prevent us from holding reference to tensors forever. This is a safety net, rather than a service guarantee, see discussion [here](https://github.com/pytorch/pytorch/issues/147168#issuecomment-2660142460).
5. Profile in async_op=False mode would look different -- collective kernels would show up in the same line and compute kernels.
Joint work with @cenzhaometa who wants to remove the event sync overhead.
Squashed contents:
* [ptd][nccl] use current-stream as nccl-stream under async=False mode (#147820)
PTD current workflow:
- PTD creates its own dedicated `ncclStream` for comm operation
- it will first add a dependency on current-stream (typically the compute stream) to ensure tensors are ready before invoking collective
such stream synchronization become expensive in Inference world (cpu overhead: 70us vs GPU kernel time: 160us).
This diff:
- async=False [default], will use current-stream as nccl-stream and avoid the stream-sync overhead
- async=True, will retain existing logic: create new nccl-stream, let it wait on current-stream to ensure tensors are ready
- pass down async from c10d down to NCCL-PG
this helps shave off 50% CPU overhead **(70us -> 35us)**, which reduce total CPU/GPU from **230us to 195us by 15%**
* [PGNCCL] Make avoid-record-stream default
* [c10d] Add asyncOp argument to Ops
* Change python side wait
* Pass asyncOp at ProcessGroup level
* Watchdog unstashing tensors as a safety net
* Stash tensors for reduce_scatter_v and all_gather_v
Pull Request approved: https://github.com/pytorch/pytorch/pull/149753
* [c10d] Move unstashing from watchdog to main thread
Pull Request approved: https://github.com/pytorch/pytorch/pull/150079
* [PGNCCL][BE] Merge mutex into TensorShelf for encapsulation
Pull Request approved: https://github.com/pytorch/pytorch/pull/150130
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150398
Approved by: https://github.com/atalman
This PR has multiple changes to `ProcessGroupNCCL` (which unfortunately are related):
1. When async_op=False, we directly launch the collective on "current" stream, instead of a trampoline stream and join back.
- Resolves#147729
- Resolves#146881
- Also saves two event syncs (which have overhead in case of HIP) and one pybind when we call `work.wait()` in distributed_c10d.py on behalf of user.
2. Entirely remove `record_stream` and use CPU-side stashing for managing tensor lifetime against recycling.
- Resolves#147168
3. Remove tensor life management when async_op=False; only use it when async_op=True.
4. To guard against user not calling `work.wait()`, we ask watchdog to unstash tensors after detecting completion of collectives, to prevent us from holding reference to tensors forever. This is a safety net, rather than a service guarantee, see discussion [here](https://github.com/pytorch/pytorch/issues/147168#issuecomment-2660142460).
5. Profile in async_op=False mode would look different -- collective kernels would show up in the same line and compute kernels.
Joint work with @cenzhaometa who wants to remove the event sync overhead.
Cc: @ngimel @awgu @Aidyn-A @skyw @wconstab @leonardo0lyj
Differential Revision: [D70937982](https://our.internmc.facebook.com/intern/diff/D70937982)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148590
Approved by: https://github.com/eqy, https://github.com/Aidyn-A, https://github.com/fduwjj
This PR has multiple changes to `ProcessGroupNCCL` (which unfortunately are related):
1. When async_op=False, we directly launch the collective on "current" stream, instead of a trampoline stream and join back.
- Resolves#147729
- Resolves#146881
- Also saves two event syncs (which have overhead in case of HIP) and one pybind when we call `work.wait()` in distributed_c10d.py on behalf of user.
2. Entirely remove `record_stream` and use CPU-side stashing for managing tensor lifetime against recycling.
- Resolves#147168
3. Remove tensor life management when async_op=False; only use it when async_op=True.
4. To guard against user not calling `work.wait()`, we ask watchdog to unstash tensors after detecting completion of collectives, to prevent us from holding reference to tensors forever. This is a safety net, rather than a service guarantee, see discussion [here](https://github.com/pytorch/pytorch/issues/147168#issuecomment-2660142460).
5. Profile in async_op=False mode would look different -- collective kernels would show up in the same line and compute kernels.
Joint work with @cenzhaometa who wants to remove the event sync overhead.
Cc: @ngimel @awgu @Aidyn-A @skyw @wconstab @leonardo0lyj
Differential Revision: [D70835197](https://our.internmc.facebook.com/intern/diff/D70835197)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148590
Approved by: https://github.com/eqy, https://github.com/Aidyn-A, https://github.com/fduwjj
This PR has multiple changes to `ProcessGroupNCCL` (which unfortunately are related):
1. When async_op=False, we directly launch the collective on "current" stream, instead of a trampoline stream and join back.
- Resolves#147729
- Resolves#146881
- Also saves two event syncs (which have overhead in case of HIP) and one pybind when we call `work.wait()` in distributed_c10d.py on behalf of user.
2. Entirely remove `record_stream` and use CPU-side stashing for managing tensor lifetime against recycling.
- Resolves#147168
3. Remove tensor life management when async_op=False; only use it when async_op=True.
4. To guard against user not calling `work.wait()`, we ask watchdog to unstash tensors after detecting completion of collectives, to prevent us from holding reference to tensors forever. This is a safety net, rather than a service guarantee, see discussion [here](https://github.com/pytorch/pytorch/issues/147168#issuecomment-2660142460).
5. Profile in async_op=False mode would look different -- collective kernels would show up in the same line and compute kernels.
Joint work with @cenzhaometa who wants to remove the event sync overhead.
Cc: @ngimel @awgu @Aidyn-A @skyw @wconstab @leonardo0lyj
Differential Revision: [D70835197](https://our.internmc.facebook.com/intern/diff/D70835197)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148590
Approved by: https://github.com/eqy, https://github.com/Aidyn-A, https://github.com/fduwjj
In Fusion model, users might change the state_dict keys by state_dict_hook
The load_state_dict APIs here won't call model.state_dict() so that the hooks won't be called to change the keys, causing the mismatch between fqn and state_dict keys.
The PR here suggests users to add how they would change the state_dict key prefix (they can name it, here we call "fqn_modifiers") by default
During loading state_dict, we have the prefix change during getting fqn so that they can be processed same as through state_dict hook.
For example:
There's a state_dict_hook:
```
def _state_dict_hook(self, destination, prefix, keep_vars):
"""Remove "embedding" from the original embedding in the state_dict
name. This keeps the orginal state dict name for the embedding
from before fusing with the FusionEmbedding.
[!Note] This update changes the order of the OrderedDict
"""
key = prefix + "embedding.weight"
new_key = prefix + "weight"
destination[new_key] = destination[key]
del destination[key]
```
In the dsd after this PR, we would skip "embedding." before "weight" if find the "fqn_modifiers" attribute at that module
```
def fqn_modifiers(self) -> Dict[str, str]:
return {
"weight": "embedding",
}
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/146557
Approved by: https://github.com/fegin
In Python 3.12, the error message has changed from "Can't pickle local object" to "Can't get local object".
The old regex would no longer catch the error.
This PR make it compatible with Python 3.12 and backward compatible as well.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/145945
Approved by: https://github.com/H-Huang
Motivation: Generalize unit tests so that can be executed for cuda and non cuda devices.
Chnages: There are general changes in common_dtesnor module for device type generalization so that tests can be executed on non cuda devices too.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139749
Approved by: https://github.com/kwen2501
Changes:
1. Bump `ruff` from 0.7.4 to 0.8.4
2. Change `%`-formatted strings to f-string
3. Change arguments with the `__`-prefix to positional-only arguments with the `/` separator in function signature.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143753
Approved by: https://github.com/Skylion007
Summary: `TestFunc` is annotated as `Callable[[object], object]` which represents a callable that takes a single argument of any type (`object`) and returns a value of any type (`object`). However, in reality, `TestFunc` could be any number of arguments, as a result, the corret typing should be `Callable[[...], object]` instead which represents a callable that takes any number of arguments (including zero) and returns a value of any type (`object`).
Test Plan: Contbuild & OSS CI
Differential Revision: D66463705
Pull Request resolved: https://github.com/pytorch/pytorch/pull/141513
Approved by: https://github.com/wz337, https://github.com/Skylion007
Adding `destroy_pg_upon_exit` property to allow derived Test classes to control whether auto destroy is desired.
(Otherwise, derived test classes will need to rewrite the `_run()` method, leading to duplicated code of `_run()` and if one needs to add things to `_run` in the future, more code change is needed.)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/141192
Approved by: https://github.com/wconstab
* Automatically applies ruff rule 401. Turns loops into equivalent list comprehensions which are faster and do not leak the scope of the loop variables.
* list comprehensions not only often have better typing, but are 50+% faster than for loops on overhead. They also preserve length information etc and are better for the interpreter to optimize.
* Manually went back and made mypy happy after the change.
* Also fixed style lints in files covered by flake8 but not by pyfmt
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140980
Approved by: https://github.com/justinchuby, https://github.com/malfet
Faced with an annoying string of warnings like this when running tests,
<img width="1644" alt="Screenshot 2024-11-15 at 11 23 21 AM" src="https://github.com/user-attachments/assets/91ff4e1d-3c29-4510-9a61-46e7df68a212">
My choices seem to be (1) call destroy_process_group() at the end of
each test fn, (2) do this in some wrapper, (3) do it in the base test
class.
Since tests in MultiProcessTestCase are responsible for calling
init_process_group themselves, they should also be responsible for
calling destroy (or at least method (3) would be asymmetric and may
result in double-destroy).
But it doesn't feel worth it to go add a destroy call manually to each
test, and try/except for a possible second destroy call seems like a
happy middle ground.
Note: tests that want to ensure that destroy runs cleanly can and should
still call destroy _inside_ the test, and this change does not affect
that.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140820
Approved by: https://github.com/fegin
Summary:
`with_comms()` is mostly used as a decorator with an optional input argument `eager_init`. The problem of a decorator with input argument is that it has to be used with invocation always, i.e., you have to use as `with_comms()` rather than `with_comms` which majority of the existing usages.
This diff tries to provide a solution such that we could use `with_comms`, `with_comms()`, `with_comms(eager_init=False)`, and `with_comms(eager_init=True)`.
Test Plan: Contbuild & OSS CI
Differential Revision: D65385700
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139637
Approved by: https://github.com/wz337
We should only pass the `device_id` when the backend is `nccl`. Otherwise, we would run into the following error:
```
RuntimeError: No backend for the parent process group or its backend does not support splitting
```
This also fixes test failure is not asserted when using `with_comms()` or `with_comms(eager_init=False)`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139097
Approved by: https://github.com/XilunWu
Add an optional `eager_init` flag to `with_comms`.
When `eager_init` is True and backend is `nccl`, we pass the `device_id` to `init_process_group()` for eager initialization.
Otherwise, `device_id` is still `None` and this goes through the normal lazy call.
Default for `eager_init` is False.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/138108
Approved by: https://github.com/kwen2501
### Fix 1: Throw async error during init wait
Previously we just busy wait for `ncclSuccess`, if the nonblocking init encountered error, we never report that. Added detection of async error via `ncclGetAsyncError`.
### Fix 2: Add wait after comm split
```
// After calling ncclCommSplit in non-blocking mode, we should wait for the
// source communicator to be out of ncclInProgress state.
// Reason 1:
// it's unsafe to call new operations on the parent comm while it's in
// ncclInProgress state.
// Reason 2:
// as of NCCL 2.23, the ptr value of child comm will not be filled until the
// state of parent comm is ncclSuccess. This may change in the future. See:
// https://github.com/NVIDIA/nccl/issues/1472
```
This wait does not mean the child comm is ready for use, neither does it block till that point.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/137741
Approved by: https://github.com/shuqiangzhang
