Now we check only that fabric allocation succeeded, but sometimes we fail during export or import afterwards, with no recourse. Check the full cycle before attempting to allocate memory with the fabric.
TODO: move it to c10/cuda so that it can be used from CUDACachingAllocator too
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160790
Approved by: https://github.com/Skylion007
This might cause some new DDEs on call sites that do not use is_contiguous_or_false() or sym_is_contiguous()
but want to find those call sites to handle this properly by calling is_contiguous_or_false() and not is_contiguous() explitly when appropriate.
I had to fix one issue after removing the implicit size oblivious reasoning. here is context
we defined in this https://github.com/pytorch/pytorch/pull/157472 sym_is_contiguous to be the function computing contiguity for dynamic shapes in c++. It returns a symbolic expression that represents contiguity and guaranteed not to throw a DDE.
when people call is_contiguous we do sym_is_contiguous().guard_bool()
when people call is_contiguous_or_false we do sym_is_contiguous().guard_or_false()
one issue not handled well was this path
```
c10::SymBool TensorImpl::sym_is_contiguous_custom(
at::MemoryFormat memory_format) const {
if (C10_UNLIKELY(matches_python_custom(SizesStridesPolicy::CustomStrides))) {
return pyobj_slot_.load_pyobj_interpreter()->is_contiguous(
this, memory_format);
}
return sym_is_contiguous_default(memory_format);
}
```
namely if we call sym_is_contiguous_custom but we have matches_python_custom(SizesStridesPolicy::CustomStrides) return true , then we used to call is_contiguous(this, memory_format);
This used to go through the load_pyobj_interpreter and end up calling the python is_contiguous call which used implicit size oblivious reasoning.
once we removed that implicit size oblivious reasoning, the right thing we want is to call
return pyobj_slot_.load_pyobj_interpreter()->sym_is_contiguous(this, memory_format);
otherwise we would get DDE even if the caller is doing sym_is_contiguous.
so I had to define it for pyinterpreter, and then I had to override it for nested tensors.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159197
Approved by: https://github.com/ezyang
Summary:
LLVM has a warning `-Wunreachable-code-break` which identifies `break` statements that cannot be reached. These compromise readability, are misleading, and may identify bugs. This diff removes such statements.
For questions/comments, contact r-barnes.
- If you approve of this diff, please use the "Accept & Ship" button :-)
Test Plan:
Sandcastle
Rollback Plan:
Differential Revision: D79835614
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160257
Approved by: https://github.com/Skylion007
# Motivation
This PR moves the implementation of `torch.cuda.memory._set_allocator_settings` to `torch._C._accelerator_setAllocatorSettings`.
Since the original API was intended as a temporary/internal utility, I am not exposing the new function as a public API.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/156175
Approved by: https://github.com/albanD
ghstack dependencies: #159629, #150312, #156165
# Motivation
Refactor `CUDAAllocatorConfig` to reuse `AcceleratorAllocatorConfig` and `ConfigTokenizer`. We would deprecate those option that overleap with `AcceleratorAllocatorConfig` in the following PR and keep them only for BC.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150312
Approved by: https://github.com/albanD
ghstack dependencies: #159629
# Motivation
While refactoring the caching allocator, I noticed that the `ExpandableSegment` constructor on CUDA had an unused parameter. This change removes that unused argument to avoid potential confusion.
# Additional Context
I noticed that `ExpandableSegment` is defined in cpp file, so it should be safe to make this change.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159356
Approved by: https://github.com/ngimel, https://github.com/albanD
ghstack dependencies: #159159
This PR is a big copy pasta from `c10/util/Float8*` -> `torch/headeronly/util/` which is why we are breaking PR sanity :C (sorry @albanD!).
Why is it not a clean copy paste?
- For BC reasons, we have to keep the old c10 file around so that OSS devs relying on those files can still get the same APIs
- Because we reexpose APIs that are headeronly through torch::headeronly, so there is an extra chunk of code in the new torch::headeronly files to do that.
Outside of the copy paste, I:
- changed the tests to call torch::headeronly instead of c10
- updated header_only_apis.txt
- added `// NOLINTNEXTLINE(bugprone-narrowing-conversions,cppcoreguidelines-narrowing-conversions)` to pass lint (which was previously skipped for -inl.h files)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159415
Approved by: https://github.com/albanD
This PR is a bit more involved but effectively works to drastically simplify PyObjectSlot and PyInterpreter.
1) For PyObjectSlot we now use a global pyinterpreter since there only is one. From here we change all of the call sites to rely on this assumption.
2) We also remove the "tags" of the PyInterpreter by deprecating `PyInterpreterStatus`.
For the reviewer, sadly it seems like `functorch/csrc/dim/dim.cpp` needed to get linted, so there is an unreadable amount of changes there. Fortunately, the only actual change in the file is as follows which just removes `getPyInterpreter()` from the `check_pyobj` call.
```
mpy::handle handle_from_tensor(Arena& A, TensorRef t) {
- // fast case: tensor is live in python
- std::optional<PyObject*> mb_obj =
- t->unsafeGetTensorImpl()->pyobj_slot()->check_pyobj(getPyInterpreter(), /*ignore_hermetic_tls=*/false);
- if (mb_obj.has_value() && !t->unsafeGetTensorImpl()->pyobj_slot()->owns_pyobj()) {
- return *mb_obj;
- }
- return A.autorelease(mpy::object::checked_steal(THPVariable_Wrap(*t)));
-}
-}
+ // fast case: tensor is live in python
+ std::optional<PyObject*> mb_obj =
+ t->unsafeGetTensorImpl()->pyobj_slot()->check_pyobj(
+ /*ignore_hermetic_tls=*/false);
+ if (mb_obj.has_value() &&
+ !t->unsafeGetTensorImpl()->pyobj_slot()->owns_pyobj()) {
+ return *mb_obj;
+ }
+ return A.autorelease(mpy::object::checked_steal(THPVariable_Wrap(*t)));
+}
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158427
Approved by: https://github.com/albanD
Essence of this copypasta:
- combine Half-inl.h and Half.h in c10/util -> torch/headeronly/util/Half.h
- Add NOLINTNEXTLINE's to the portions of Half-inl.h that were previously in the ignore list of clangtidy
- Re-expose all APIs in namespaces and through includes of the original files. Ideally, we would have the APIs in torch::headeronly and reexpose them in c10, but that runs into BC issues (see D78997465) so for now we are keeping the APIs in c10 but reexposing them in torch::headeronly.
- Change test cases in test_aoti_abi_check to test torch::headeronly::Half vs c10::Half (they're the same thing but we eventually want all the tests for headeronly APIs to only import from headeronly).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159172
Approved by: https://github.com/albanD, https://github.com/desertfire
# Motivation
This PR moves the implementation of `torch.cuda.memory._set_allocator_settings` to `torch._C._accelerator_setAllocatorSettings`.
Since the original API was intended as a temporary/internal utility, I am not exposing the new function as a public API.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/156175
Approved by: https://github.com/albanD
ghstack dependencies: #149601, #157908, #150312, #156165
# Motivation
Refactor `CUDAAllocatorConfig` to reuse `AcceleratorAllocatorConfig` and `ConfigTokenizer`. We would deprecate those option that overleap with `AcceleratorAllocatorConfig` in the following PR and keep them only for BC.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150312
Approved by: https://github.com/albanD
ghstack dependencies: #149601, #157908
# Motivation
This PR aims to generalize `AllocatorConfig` to be device-agnostic. Introduce the class `AcceleratorAllocatorConfig` to clarify its scope as a configuration manager for accelerator backends (e.g., CUDA, XPU). The another name `AllocatorConfig` is now reserved for a potential future base class that can unify configuration handling for both CPU and accelerator allocators, should similar requirements arise for the CPU path.
# Design Rule
## Overall
This class configures memory allocation for both device and host memory. A single `AcceleratorAllocatorConfig` instance is shared across all accelerator backends, such as CUDA and XPU, under the assumption that relevant environment variables apply uniformly to all accelerators. Device-specific configuration extensions are supported via hooks (see `registerDeviceConfigParserHook`).
Introduce a new class `ConfigTokenizer` to help process the env variable config key-value pair
## Naming Convention:
- Public API names in `AcceleratorAllocatorConfig` should be device-generic.
- Members prefixed with `pinned_` are specific to the host/pinned allocator.
- Environment variable names should be generic across backends.
- Comma-separated key-value pairs in the format: `key:value`. Use square brackets `[]` for list values Example: `key1:123, key2:[val1,val2]`
## Environment Variables:
- The default environment variable for configuration is `PYTORCH_ALLOC_CONF`.
- For backward compatibility, `PYTORCH_CUDA_ALLOC_CONF` and `PYTORCH_HIP_ALLOC_CONF` are also supported with lower priority.
Differential Revision: [D79011786](https://our.internmc.facebook.com/intern/diff/D79011786)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/149601
Approved by: https://github.com/albanD
Essence of this copypasta:
- combine Half-inl.h and Half.h in c10/util -> torch/headeronly/util/Half.h
- Add NOLINTNEXTLINE's to the portions of Half-inl.h that were previously in the ignore list of clangtidy
- Re-expose all APIs in namespaces and through includes of the original files. Ideally, we would have the APIs in torch::headeronly and reexpose them in c10, but that runs into BC issues (see D78997465) so for now we are keeping the APIs in c10 but reexposing them in torch::headeronly.
- Change test cases in test_aoti_abi_check to test torch::headeronly::Half vs c10::Half (they're the same thing but we eventually want all the tests for headeronly APIs to only import from headeronly).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159172
Approved by: https://github.com/albanD, https://github.com/desertfire
Straightup copy pasta. Keeps APIs in c10 and reexposes them to torch::headeronly.
It is arguable that we should just get rid of some of these unused dtypes but that is outside the scope of this PR, which is meant to build up to ScalarType moving to headeronly.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159302
Approved by: https://github.com/malfet, https://github.com/albanD
# Motivation
While refactoring the caching allocator, I noticed that the `AllocParams` constructor on CUDA had an unused parameter. This change removes that unused argument to avoid potential confusion.
# Additional Context
I noticed that `AllocParams` is defined in cpp file, so it should be safe to make this change.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159159
Approved by: https://github.com/cyyever, https://github.com/albanD
This PR makes some less risky changes to PyObjectSlot as there is a lot of stuff we do not need since there is only one interpreter. Specifically `check_interpreter` and `has_pyobj_nonhermetic` are removed
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158407
Approved by: https://github.com/albanD
ghstack dependencies: #158290, #158291
