----
# Refactor and Improve the OpenReg Module
## Background
Since PrivateUse1 has become the main path for integrating new devices with PyTorch, there have been some feature requests related to PrivateUse1 regarding interfaces, documentation, reference examples, etc., such as the following:
- https://github.com/pytorch/pytorch/issues/155864
- https://github.com/pytorch/pytorch/issues/144955
- https://github.com/pytorch/pytorch/issues/144845
Taking these requests into consideration and combining them with the position of OpenReg, which is currently used as the test backend for PrivateUse1, I'm planning to make the following optimizations:
- Optimize the implementation of OpenReg to make it align with the standard specifications for real backend (C++) access, serving as a reference for new device integration code.
- Add comprehensive documentation to the [developer notes](https://docs.pytorch.org/docs/main/notes.html) to guide new accelerator integration, functioning as a reference manual.
## Design Principles:
- Minimization Principle: Keep the code small and clear; only implement the minimum set of code required for verification and as an integration reference.
- Authenticity Principle: Integrate OpenReg in the same way that real accelerators access PyTorch.
## More Infos:
Pleaes refer to [this](6b8020f1ab/test/cpp_extensions/open_registration_extension/torch_openreg/README.md) for more information about `OpenReg`.
## Current Progress:
- Refer to the implementation of [torch_xla](https://github.com/pytorch/xla) to refactor all of OpenReg's code, making it easier to understand.
- Ensure all tests in [test/test_openreg.py](https://github.com/FFFrog/pytorch/blob/openreg/test/test_openreg.py) pass after refactoring.
## Next Steps:
- Add more features to cover all integration points.
- Gradually add user guides and documentation to the [developer notes](https://docs.pytorch.org/docs/main/notes.html).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158090
Approved by: https://github.com/seemethere, https://github.com/albanD
```
// The torch::stable::Tensor class is a highlevel C++ header-only wrapper around
// the C shim Tensor APIs. We've modeled this class after TensorBase, as custom
// op kernels only really need to interact with Tensor metadata (think sizes,
// strides, device, dtype). Other functions on Tensor (like empty_like) should
// live like the ATen op that they are and exist outside of this struct.
//
// There are several goals of this class over AtenTensorHandle and
// RAIIAtenTensorHandle:
// 1. torch::stable::Tensor is a nicer UX much closer to torch::Tensor than the
// C APIs with AtenTensorHandle. Under the hood we still call to these C shim
// APIs to preserve stability.
// 2. RAIIAtenTensorHandle boils down to a uniq_ptr that forces the user to pass
// around ownership. This makes it difficult to pass one input into 2
// different functions, e.g., doing something like c = a(t) + b(t) for
// stable::Tensor t. Thus, we use a shared_ptr here.
```
This PR:
- exemplifies the above
- adds test cases in libtorch_agnostic to make sure the file actually works
- includes the results of a battle with template specialization
Pull Request resolved: https://github.com/pytorch/pytorch/pull/155367
Approved by: https://github.com/albanD
This test was never the shining star in class but it helped check that we can properly delete a stable library. But now that we are running it in CI this is not a good test to annoy people with as dlclose + parallelism is likely not the move. I will miss it locally though.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153975
Approved by: https://github.com/jbschlosser
Related: #148920
This PR:
* Introduces a new file `test/cpp_extensions/python_agnostic_extension/test/test_python_agnostic.py` with testing that follows the usual python testing patterns
* This replaces the testing for python_agnostic in `test/test_cpp_extensions_aot.py`
After this PR, it is now possible to run:
```
python test/cpp_extensions/python_agnostic_extension/test/test_python_agnostic.py
```
and the test will build the prerequisite wheel before running the tests.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153274
Approved by: https://github.com/janeyx99, https://github.com/cyyever
ghstack dependencies: #153264
Related: #148920
This PR:
* Provides a helper `install_cpp_extension(extension_root)` for building C++ extensions. This is intended to be used in `TestMyCppExtension.setUpClass()`
* Updates libtorch_agnostic tests to use this
* Deletes preexisting libtorch_agnostic tests from `test/test_cpp_extensions_aot.py`
* Fixes `run_test.py` to actually run tests in `test/cpp_extensions/libtorch_agnostic_extension/test/test_libtorch_agnostic.py` to avoid losing coverage. This wasn't being run due to logic excluding tests that start with "cpp"; this is fixed now
After this PR, it is now possible to run:
```
python test/cpp_extensions/libtorch_agnostic_extension/test/test_libtorch_agnostic.py
```
and the test will build the `libtorch_agnostic` extension before running the tests.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153264
Approved by: https://github.com/janeyx99
As the title stated.
**Changes:**
- Remove unnecessary header file
- Remove unnecessary registry logic about PrivateUse1HooksRegistry，such as TORCH_DECLARE_REGISTRY, C10_DEFINE_REGISTRY, etc,.
- using static + global variable to do initialization instead of call_one
**Next Step:**
Enable test_openreg.py in CI/CD to guard the quality of PrivateUse1
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151005
Approved by: https://github.com/albanD
I want to format and refactor the csrc file of pytorch_openreg. To make the code review clearer and easier to understand, I divide the code refactoring into two parts:
- Part 1: Code formatting
- Part 2: Code refactoring and optimization (Next PR)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151004
Approved by: https://github.com/albanD
ghstack dependencies: #151000
This change does 2 important things:
(a) Instead of relying on IValue type as source of truth, we use the schema as the source of truth, which is important as IValue types are overloaded and can ambiguously convert incorrectly. For example, a MemoryFormat will look like an int + get converted to an int64_t vs a MemoryFormat!
(b) This PR expands support for many more types to encompass way more schemas, e.g., Optional, Device, dtype, etc. The main win from this PR is the ability for aoti_torch_call_dispatcher to call TensorFactory ops like ones_like/empty_like!
Pull Request resolved: https://github.com/pytorch/pytorch/pull/149052
Approved by: https://github.com/albanD
This PR adds two main parts:
- shim.h stable C APIs into torch::Library APIs
- a higher level API in torch/csrc/stable/library.h that calls into this shim.h + otherwise is self contained
Goal: custom kernel writers should be able to call the apis in the directories above in order to register their library in a way that allows their custom extension to run with a different libtorch version than it was built with.
Subplots resolved:
- Do we want a whole separate StableLibrary or do we want to freeze torch::Library and add `m.stable_impl(cstring, void (*fn)(void **, int64_t, int64_t)` into it
- Yes, we want a separate StableLibrary. We cannot freeze Library and it is NOT header only.
- Should I use unint64_t as the common denominator instead of void* to support 32bit architectures better?
- Yes, and done
- Should I add a stable `def` and `fragment` when those can be done in python?
- I think we do want these --- and now they're done
- Where should library_stable_impl.cpp live? -- no longer relevant
- I need some solid test cases to make sure everything's going ok. I've intentionally thrown in a bunch of random dtypes into the signature, but I still haven't tested returning multiple things, returning nothing, complex dtypes, etc.
- Have since tested all the torch library endpoints. the others can be tested in a followup to separate components that need to be in shim.h vs can be added later
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148124
Approved by: https://github.com/albanD, https://github.com/zou3519, https://github.com/atalman
This PR adds two main parts:
- shim.h stable C APIs into torch::Library APIs
- a higher level API in torch/csrc/stable/library.h that calls into this shim.h + otherwise is self contained
Goal: custom kernel writers should be able to call the apis in the directories above in order to register their library in a way that allows their custom extension to run with a different libtorch version than it was built with.
Subplots resolved:
- Do we want a whole separate StableLibrary or do we want to freeze torch::Library and add `m.stable_impl(cstring, void (*fn)(void **, int64_t, int64_t)` into it
- Yes, we want a separate StableLibrary. We cannot freeze Library and it is NOT header only.
- Should I use unint64_t as the common denominator instead of void* to support 32bit architectures better?
- Yes, and done
- Should I add a stable `def` and `fragment` when those can be done in python?
- I think we do want these --- and now they're done
- Where should library_stable_impl.cpp live? -- no longer relevant
- I need some solid test cases to make sure everything's going ok. I've intentionally thrown in a bunch of random dtypes into the signature, but I still haven't tested returning multiple things, returning nothing, complex dtypes, etc.
- Have since tested all the torch library endpoints. the others can be tested in a followup to separate components that need to be in shim.h vs can be added later
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148124
Approved by: https://github.com/albanD, https://github.com/zou3519
This patch adds support for sycl kernels build via `torch.utils.cpp_extension.load`, `torch.utils.cpp_extension.load_inline` and (new) `class SyclExtension` APIs. Files having `.sycl` extension are considered to have sycl kernels and are compiled with `icpx` (dpc++ sycl compiler from Intel). Files with other extensions, `.cpp`, `.cu`, are handled as before. API supports building sycl along with other file types into single extension.
Note that `.sycl` file extension is a PyTorch convention for files containing sycl code which I propose to adopt. We did follow up with compiler team to introduce such file extension in the compiler, but they are opposed to this. At the same time discussion around sycl file extension and adding sycl language support into such tools as cmake is ongoing. Eventually cmake also considers to introduce some file extension convention for sycl. I hope we can further influence cmake and compiler communities to broader adopt `.sycl` file extension.
By default SYCL kernels are compiled for all Intel GPU devices for which pytorch native aten SYCL kernels are compiled. At the moment `pvc,xe-lpg`. This behavior can be overridden by setting `TORCH_XPU_ARCH_LIST` environment variables to the comma separated list of desired devices to compile for.
Fixes: #132944
CC: @gujinghui @EikanWang @fengyuan14 @guangyey @jgong5
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132945
Approved by: https://github.com/albanD, https://github.com/guangyey, https://github.com/malfet
Co-authored-by: Nikita Shulga <2453524+malfet@users.noreply.github.com>
This patch adds support for sycl kernels build via `torch.utils.cpp_extension.load`, `torch.utils.cpp_extension.load_inline` and (new) `class SyclExtension` APIs. Files having `.sycl` extension are considered to have sycl kernels and are compiled with `icpx` (dpc++ sycl compiler from Intel). Files with other extensions, `.cpp`, `.cu`, are handled as before. API supports building sycl along with other file types into single extension.
Note that `.sycl` file extension is a PyTorch convention for files containing sycl code which I propose to adopt. We did follow up with compiler team to introduce such file extension in the compiler, but they are opposed to this. At the same time discussion around sycl file extension and adding sycl language support into such tools as cmake is ongoing. Eventually cmake also considers to introduce some file extension convention for sycl. I hope we can further influence cmake and compiler communities to broader adopt `.sycl` file extension.
By default SYCL kernels are compiled for all Intel GPU devices for which pytorch native aten SYCL kernels are compiled. At the moment `pvc,xe-lpg`. This behavior can be overridden by setting `TORCH_XPU_ARCH_LIST` environment variables to the comma separated list of desired devices to compile for.
Fixes: #132944
CC: @gujinghui @EikanWang @fengyuan14 @guangyey @jgong5
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132945
Approved by: https://github.com/albanD, https://github.com/guangyey
This could be BC breaking, because there was a period of time when we use py_limited_api=True but don't enforce the flag, and now that we will start enforcing the flag, people's custom extensions may fail to build.
This is strictly still better behavior, as it is sketchy to claim CPython agnosticism without the flag, but calling this out as potential people yelling at us. Ways to mitigate this risk + reasons this may not be too big a deal:
- People haven't known about py_limited_api for extensions much due to lack of docs from python so usage is low right now
- My current tutorial is in store to make new users of py_limited_api pass this flag, so it'd be a noop for them.
Test plan:
* Locally i'm confident as I tried rebuilding ao with this change and it reliably failed (cuz importing torch/extension.h is a nono)
* Unit test wise, the normal python_agnostic one I added should work
Pull Request resolved: https://github.com/pytorch/pytorch/pull/145764
Approved by: https://github.com/ezyang, https://github.com/zou3519, https://github.com/albanD
