Previously, when we applied a replacement, a SymInt that was
previously an unbacked SymInt would then transmute into whatever
we replaced it into (e.g., a constant).
This has a major downside: we often look at SymInts associated with
FX nodes (e.g., the meta of x.item() return) to find out where the
unbacked SymInt was allocated. If we replace it, we no longer can
find out where, e.g., u1 was allocated! But we need to know this
so we can generate deferred runtime asserts like u1 == s0.
To solve this problem, I have a special mode for replace, resolve_unbacked=False, which lets you disable substitutions on unbacked SymInts. When reporting node.expr, we preferentially avoid applying unbacked SymInt substitutions. To understand if we might accidentally reapply the substitution later, before we have reached the deferred runtime assert, we must study the calls to simplify() in ShapeEnv. My audit turns up these sites:
* `produce_guards`: this is fine, deferred runtime asserts never show up here, we must NOT have unbacked SymInts show up here. Similarly `get_nontrivial_guards`.
* `_maybe_evaluate_static`: this is fine, we are using this to determine if it is necessary to produce a guard/runtime assert. We don't want to reissue a runtime assert if we've already asserted on it, and replacements can help us understand if this has occurred.
* `_simplify_floor_div`: this is a legitimate bug, it needs to be `resolve_unbacked=False`
* `_refine_ranges`: this is fine, a refined range doesn't affect what runtime asserts we issue
* `_update_divisible`: this updates the `self.divisible` set, which specifies when we can simplify away divisibility constraints. Since this affects replacements only, it won't cause us to oversimplify a user provided expression.
There are some situations where we DO want to always apply the substitution, specifically when we have the duplicate symbol problem (we retrace an item call and get u0 and u1 which refer to the same thing.) I don't want two symbols in this case, so a special `rename_unbacked_to` is provided which sets up the unconditional renaming.
Along the way, I make a refinement to `_update_var_to_range`: if you update a var range for a size-like unbacked SymInt, you are now no longer allowed to set its lower bound below 2. This is because if you could, then our size oblivious tests for it would be inconsistent. Actually, I think there is still some inconsistency, because if you assert `u0 == 0` we will still end up with this in deferred runtime asserts, and we will then use this to simplify these statements to be True everywhere else. Maybe we should forbid this kind of refinement; not done in this PR.
Fixes https://github.com/pytorch/pytorch/issues/119689
Fixes https://github.com/pytorch/pytorch/issues/118385
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/120816
Approved by: https://github.com/lezcano
See #113541
The PR allows for registering and controlling multiple RNG states using indices, ensuring cudagraph-safe operations, and includes both C++ and Python API changes to support this functionality.
cc @eellison @anijain2305 @jansel @ezyang @ptrblck @csarofeen @mcarilli
Pull Request resolved: https://github.com/pytorch/pytorch/pull/114068
Approved by: https://github.com/ezyang
This PR corrects the example in the AOTInductor example which currently fails with:
```
/home/ubuntu/test/inference.cpp:21:62: error: cannot bind non-const lvalue reference of type ‘std::vector<at::Tensor>&’ to an rvalue of type ‘std::vector<at::Tensor>’
21 | std::cout << runner.run({torch::randn({2, 10}, at::kCPU)})[0] << std::endl;
|
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/121672
Approved by: https://github.com/desertfire
Summary:
## No Functional Change
- Refactor Subprocess Handler into a separate folder for easier subclassing
- SubprocessHandler
- added `local_rank_id` in `SubprocessHandler` to make it available as a field in the class
- pass in `local_rank_id` from subprocess start
Test Plan: No functional changes.
Differential Revision: D54038627
#suppress-api-compatibility-check
Pull Request resolved: https://github.com/pytorch/pytorch/pull/120373
Approved by: https://github.com/kurman
As titled, this PR introduces a dedicated `ParallelStyle` to shard the
nn.LayerNorm/nn.Dropout/RMSNorm layers. We were mainly using a manual
distribute_module calls before when sharding the RMSNorm layer, but I
think we should have a dedicate TP API to easily shard those layers,
instead of user manually using DTensors.
I call this SequenceParallel, which might bring some confusion that we
technically "deprecated" a SequenceParallel style months ago. But this
time the SeuqenceParallel style is significantly different with the
previous ones (which used to shard two consecutive Linear layers). I
believe making it the right name is the first priority, instead of
worrying about the issue of reusing the old name
Pull Request resolved: https://github.com/pytorch/pytorch/pull/121295
Approved by: https://github.com/awgu, https://github.com/tianyu-l
ghstack dependencies: #121294
Summary: WrapperModule seems a good idea but may introduce some surprising behavior to users, for example, it never registers enclosed modules as submodules and therefore it's unclear that's the state dict for the exported program should look like, because some people may argue to include every state in state dict but others want to keep them as constants.
Test Plan: CI
Reviewed By: tugsbayasgalan
Differential Revision: D54326331
Pull Request resolved: https://github.com/pytorch/pytorch/pull/121042
Approved by: https://github.com/angelayi
Loss parallel is the last piece of sequence parallelism to enable. It enables efficient distributed cross entropy computation when the input is sharded on the class dimension (in a classification problem with many classes). The implementation is via a context manager `loss_parallel`, after enabling which users can directly use `torch.nn.functional.cross_entropy` or `torch.nn.CrossEntropyLoss` without modifying other parts of their code.
Here are the underlying rationales why we are going through these op replacements:
1. `nn.functional.cross_entropy` is the common method that OSS user is using for things like transformer training, to avoid changing user code, we want user to still use this function for loss calculation if they are already using it.
2. `nn.functional.cross_entropy` boils down into `aten.log_softmax` and `aten.nll_loss_foward/backward`, and DTensor now supports those ops already (#117723#119255#118917#119256). They are doing computation with input *replicated* on the class dimension.
3. However when the input of this loss calculation is **sharded on the class dimension**, to run sharded computation efficiently, we need to run both `aten.log_softmax` and `aten.nll_loss_foward` with multiple all-reduce collectives **in the middle of** those aten ops. This is not possible if we are just overriding these two ops, so we need to have some way to **decompose** these two ops into smaller ops to have collectives run in the middle of these two ops.
4. We explored the existing decompositions (#118950). It seems working, except that `log_softmax_backward` and `nll_loss_backward` combined together in aten are implemented in a inefficient way, which would trigger an additional expensive collective. Recently some user also reported similar issues https://github.com/pytorch/pytorch/issues/119261.
5. Therefore, currently we are doing our own decomposition inside a context manager for sequence parallelism specifically. Once we have a better decomposition in core, we can possibly take that instead of reinventing the wheels here.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/119877
Approved by: https://github.com/wanchaol
Summary:
Pulling out logging parameters into a logging specs that can be overridden (follow-up changes on possible mechanism)
Why?
Right now the logging approach is quite rigid:
- Requires for log directory to exist and not be empty
- Will create tempdir otherwise,
- Creates subdir for a run
- creates subdir for each attempt
- creates files named as stdout.log, stderr.log, error.json
In some instances some of the users would like to customize the behavior including file names based on context. And we do have right now a mechanism to template multiplexed teed output prefix.
With current changes, users can create custom log spec that can use env variables to change the behavior.
Notes:
Made `LaunchConf.logs_specs` as an optional field that will be bound to `DefaultLogsSpecs` instance. There are large number of clients (code) that use the API directly without using torchrun API. For those cases, we have to explicitly pass LogSpecs implementation if we would like to override the implementation. For the regular torchrun users, we can use pluggable approach proposed in the follow up change.
Test Plan: CI + unit tests
Differential Revision: D54176265
Pull Request resolved: https://github.com/pytorch/pytorch/pull/120691
Approved by: https://github.com/ezyang
Summary: In non-strict mode of torch.export() we didn't set those `is_compiling()` to `True` which is needed by some models.
Test Plan: Unit tests and manual testing.
Differential Revision: D53624452
Pull Request resolved: https://github.com/pytorch/pytorch/pull/119602
Approved by: https://github.com/suo
This PR adds a couple of missing words in the Checkpointing documentation, it doesn't have a specific issue number related to it.
Changes are:
- "backward." -> "backward propagation."
- "to be advanced than" -> "to be more advanced than"
Pull Request resolved: https://github.com/pytorch/pytorch/pull/120196
Approved by: https://github.com/soulitzer
Summary: This commit adds the `model_is_exported` util function
for users to be able to easily tell what APIs to call to move
their models between train and eval modes. This has the
additional advantage of hiding the implementation of how we
detect a model is exported, in case the metadata format changes
in the future.
Test Plan:
python test/test_quantization.py TestQuantizePT2E.test_model_is_exported
Differential Revision: [D53812972](https://our.internmc.facebook.com/intern/diff/D53812972)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/119726
Approved by: https://github.com/tugsbayasgalan, https://github.com/albanD
# Motivation
According to [[RFC] Intel GPU Runtime Upstreaming](https://github.com/pytorch/pytorch/issues/114842) and [[RFC] Intel GPU Runtime Upstreaming for Allocator](https://github.com/pytorch/pytorch/issues/116322), we will upstream the key functionality of device `Allocator` dedicated for XPU to PyTorch. And following our design prepare to generalize `Allocator` in parallel.
# Design
In the current design, XPU uses an `XPUAllocator` class, inherited from `c10::Allocator`. `XPUAllocator` is a manager to handle `DeviceCachingAllocator`, which is a per-device implementation of the caching mechanism to manage the already cached or newly allocated memory. The caching mechanism is similar to other backends, like CUDA. We can visualize the design as below.
<p align="center">
<img width="162" alt="image" src="https://github.com/pytorch/pytorch/assets/106960996/6b17b8cf-e7d1-48b4-b684-f830c409d218">
</p>
# Additional Context
We're going to implement our design gradually. This PR covers the device `Allocator` dedicated to XPU. The second PR covers the host `Allocator`.
Besides these PRs, we plan to generalize the device `Allocator` device-agnostic through another PR.
In this PR, our device `Allocator` has the same memory management mechanism as CUDA, but lacks features such as expendable segments and statistics. We will add these features back in the subsequent PR which intend to generalize `Allocator`.
The differences with CUDA:
only key functionality, and lack of AsyncAllocator, gpu_trace, history_record, graph functionality, memory snapshot, memory statistics, expandable segment...
Pull Request resolved: https://github.com/pytorch/pytorch/pull/118091
Approved by: https://github.com/EikanWang, https://github.com/gujinghui, https://github.com/jgong5, https://github.com/albanD
ghstack dependencies: #117611, #117619, #117734
# Motivation
As mentioned in [[RFC] Intel GPU Runtime Upstreaming](https://github.com/pytorch/pytorch/issues/114842), the next runtime component we would like to upstream is `Event` which handles the status of an operation that is being executed. Typically, in some circumstances, we can fine-grain control of the operation execution via `Event`.
# Design
`XPUEvent` is a movable but not a copyable wrapper around sycl event. It should be created lazily on an XPU device when recording an `XPUStream`. Meanwhile, `XPUEvent` can wait for another `XPUEvent` or all the submitted kernels on an `XPUStream` to complete. Align to the other backend, the C++ files related to `Event` will be placed in `aten/src/ATen/xpu` folder. For frontend code, `XPUEvent` runtime API will be bound to Python `torch.xpu.Event`. The corresponding C++ code will be placed in `torch/csrc/xpu/Event.cpp` and Python code will be placed in `torch/xpu/streams.py` respectively.
# Additional Context
It is worth mentioning that the `elapsed_time` method is temporarily not supported by `XPUEvent`. We will be adding support for it soon. Meanwhile `XPUEvent` doesn't support IPC from different processes. For the other parts, we have almost a 1:1 mapping with CUDA.
lack of the below APIs:
- `torch.cuda.Event.ipc_handle`
- `CUDAEvent`'s constructor with `IpcEventHandle`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/117734
Approved by: https://github.com/EikanWang, https://github.com/gujinghui, https://github.com/jgong5, https://github.com/malfet
ghstack dependencies: #117611, #117619
# Motivation
According to [[1/2] Intel GPU Runtime Upstreaming for Stream](https://github.com/pytorch/pytorch/pull/117611), as mentioned in [[RFC] Intel GPU Runtime Upstreaming](https://github.com/pytorch/pytorch/issues/114842), the second PR covers the changes under `python frontend`.
# Design
Currently, it primarily offers stream-related APIs, including
- `torch.xpu.StreamContext`
- `torch.xpu.current_stream`
- `torch.xpu.set_stream`
- `torch.xpu.synchronize`
- `torch._C._xpu_getCurrentRawStream`
# Additional Context
We will implement functions like `torch.xpu.Stream.wait_event`, `torch.xpu.Stream.wait_stream`, and `torch.xpu.Stream.record_event` in the next PR related with `Event`.
The differences with CUDA:
no default and external stream in XPU and lack of below APIs:
- `torch.cuda.ExternalStream`
- `torch.cuda.default_stream`
- `toch.cuda.is_current_stream_capturing`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/117619
Approved by: https://github.com/EikanWang, https://github.com/jgong5, https://github.com/gujinghui, https://github.com/albanD
ghstack dependencies: #117611
Added a `torch.Tensor` method that defines how to transform `other`, a value in the state dictionary, to be loaded into `self`, a param/buffer in an `nn.Module` before swapping via `torch.utils.swap_tensors`
* `param.module_load(sd[key])`
This method can be overridden using `__torch_function__`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/117913
Approved by: https://github.com/albanD
# Motivation
According to [[1/4] Intel GPU Runtime Upstreaming for Device](https://github.com/pytorch/pytorch/pull/116019), as mentioned in [[RFC] Intel GPU Runtime Upstreaming](https://github.com/pytorch/pytorch/issues/114842), this last PR covers the changes under lazy initialization.
# Design
This PR primarily offers the support of multi-processing via lazy initialization. We lazily initialize our runtime avoiding initializing XPU until the first time it is accessed. In our design, we extend `cuda_lazy_init` to `device_lazy_init` which is a device-agnostic API that can support any backend. And change `maybe_initialize_cuda` to `maybe_initialize_device` to support lazy initialization for both CUDA and XPU while maintaining scalability.
# Additional Context
We adopt a similar design to CUDA. So we share some code with CUDA.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/116869
Approved by: https://github.com/EikanWang, https://github.com/jgong5, https://github.com/gujinghui, https://github.com/malfet
ghstack dependencies: #119248
Added `torch.__future__.{get/set}_swap_module_params_on_conversion` that defaults to `False` for now, but we probably want to modify to override this and default to `True` in `nn.Module._apply` if input is a tensor subclass.
From offline discussion, for now we are **not** allowing `swap_tensor` after the first module forward has been run*** if the autograd graph is still alive. The reason being that `torch.utils.swap_tensors(t1, t2)` requires the `use_count` of both `TensorImpl`s associated with `t1` and `t2` to be 1. The first forward pass will install `AccumulateGrad` nodes on each param, which [bump the refcount of the associated TensorImpl](6cf1fc66e3/torch/csrc/autograd/variable.cpp (L307)). **Future work might be to swap the refs that the `AccumulateGrad` nodes hold if it is necessary.**
***From this, it might seem like we don't need to handle gradients. However, I still handle the grads for the edge case that the grads are set via `p.grad = grad` OR the autograd graph is no longer alive because the output has been garbage collected.
If any `swap_tensors` fails on any of the parameters in the `nn.Module` we raise an error.
**`RNNBase` overrides `nn.Module._apply()` and installs weakrefs on some parameters. As a result, all modules that inherit from `RNNBase` (`RNN`, `GRU` and `LSTM`) cannot use the`swap_tensors` path as of now**
Pull Request resolved: https://github.com/pytorch/pytorch/pull/117167
Approved by: https://github.com/albanD
ghstack dependencies: #118028
