In this PR, we are implementing Functionalization on pre-dispatch graph. Today, every dispatch key except for Dispatchkey.Python has a dedicated mode stack in python. PreDispatch tracing relies on this behaviour by pushing ProxyTorchDispatchMode to Dispatchkey.PreDispatch mode stack and handle the dispatching logic in python. To make pre-dispatch functionalization work, we now need to push FunctionalTensorMode on DispatchKey.PreDispatch mode stack and make sure it runs before ProxyTorchDispatchMode. (this is very similar to how post-dispatch tracing work). Here are some design decisions we made for this flow to work:
1. FunctionalTensorMode internally calls C++ functionalize key. Since C++ functionalization goes after PreDispatch, if we are not careful, we will keep re-entering into PreDispatch key. We solve this by directly dispatching to C++ Functionalize key.
2. We delete mode_stack_per_key logic because the only realistic time it is exercised is for PreDispatch and it is in general not safe to have a plain list because FunctionalTensorMode and ProxyTorchDispatchMode ordering matter and it is hard to enforce it on plain list. Instead, now we have a private class that tracks PreDispatch mode stack.
3. We will still run CompositeImplicitAutograd decomps in this PR, and disable this logic later as a followup.
Some missing bits after this PR:
1. Preserving autograd ops in a functional form. Right now they still show up in the graph but in a "non-functional" way.
2. Turn off CompositeImplicitAutograd decomps
3. Functionalizing HOO
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113728
Approved by: https://github.com/bdhirsh
In certain edge cases when using lazy tensors, the base tensor stored in the `FunctionalStorageImpl` and the `value_` tensor stored in the `FunctionalTensorWrapper` diverge. For instance, take this simple example
```python
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
self.fc1 = torch.nn.Linear(4, 2, bias=False)
def forward(self, x):
return x @ self.fc1.weight.transpose(0, 1)
with torch.device("lazy"):
model = Model()
x = torch.ones(4)
out = model(x)
```
The call to `transpose` on the lazily initialized weight `fc1.weight` applies a view op on the functional tensor which only gets propagated to the functional tensor wrapper and not the base tensor in the storage. Thus, causing them to diverge.
To fix this behaviour, we need to reset the functional tensor's storage. To facilitate this, we add a `reset_storage` method to `FunctionalTensorWrapper` which clears away the old storage and view metas.
CC: @behzad-a @GlebKazantaev @wconstab @bdhirsh
Pull Request resolved: https://github.com/pytorch/pytorch/pull/115235
Approved by: https://github.com/bdhirsh
This is the cheap and cheerful implementation, which is only enabled on TORCH_SHOW_CPP_STACKTRACES, because it *eagerly* symbolizes immediately at exception throw time, even if the exception will end up getting caught. It would be better to do this lazily and only symbolize when we try to print the exception, but that requires a more involved refactor of c10::Error that I don't feel like doing.
Compare the output before:
```
frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x95 (0x7fa21b99d975 in /data/users/ezyang/c/pytorch/torch/lib/libc10.so)
frame #1: c10::TensorImpl::throw_cannot_call_with_symbolic(char const*) const + 0x8d (0x7fa21b951269 in /data/users/ezyang/c/pytorch/torch/lib/libc10.so)
frame #2: c10::TensorImpl::sizes_custom() const + 0x9f (0x7fa21b9770df in /data/users/ezyang/c/pytorch/torch/lib/libc10.so)
frame #3: at::meta::structured_mm::meta(at::Tensor const&, at::Tensor const&) + 0x31e (0x7fa20a202a8e in /data/users/ezyang/c/pytorch/torch/lib/libtorch_cpu.so)
frame #4: <unknown function> + 0x29f34de (0x7fa20b5f34de in /data/users/ezyang/c/pytorch/torch/lib/libtorch_cpu.so)
frame #5: <unknown function> + 0x2a1fd8e (0x7fa20b61fd8e in /data/users/ezyang/c/pytorch/torch/lib/libtorch_cpu.so)
frame #6: <unknown function> + 0x6b907b (0x7fa2142b907b in /data/users/ezyang/c/pytorch/torch/lib/libtorch_python.so)
frame #7: <unknown function> + 0x6b6175 (0x7fa2142b6175 in /data/users/ezyang/c/pytorch/torch/lib/libtorch_python.so)
```
and after:
```
#4 c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) from ??:0
#5 c10::TensorImpl::throw_cannot_call_with_symbolic(char const*) const from ??:0
#6 c10::TensorImpl::sizes_custom() const [clone .localalias] from TensorImpl.cpp:0
#7 at::meta::structured_mm::meta(at::Tensor const&, at::Tensor const&) from ??:0
#8 at::(anonymous namespace)::wrapper_Meta_mm_out_out(at::Tensor const&, at::Tensor const&, at::Tensor&) from RegisterMeta.cpp:0
#9 c10::impl::make_boxed_from_unboxed_functor<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor& (at::Tensor const&, at::Tensor const&, at::Tensor&), &at::(anonymous namespace)::wrapper_Meta_mm_out_out>, at::Tensor&, c10::guts::typelist::typelist<at::Tensor const&, at::Tensor const&, at::Tensor&> >, false>::call(c10::OperatorKernel*, c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) from RegisterMeta.cpp:0
```
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113207
Approved by: https://github.com/Skylion007
Summary:
We've made the following changes:
- The new way to use the API is `m.impl_abstract_pystub(module, context)`.
Every subsequent m.def of an op inside the TORCH_LIBRARY block gives
the op the `impl_abstract_pystub`.
- Added a mechanism to determine if an operator was defined in Python or C++.
Library.define in Python appends the op to a global set, which is analogous
to what we do for tracking Library.impl.
- If someone does `torch.library.impl_abstract` in Python for an operator, then
we require that it has an `impl_abstract_pystub` specified and we also check
that the module in the `impl_abstract_pystub` is the same as the module where
the call to `torch.library.impl_abstract` exists.
- Unfortunately we can't check the "context" (which is the buck target on
buck-based systems) because buck sits above us.
bypass-github-export-checks
Test Plan: - existing tests
Differential Revision: D51080493
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113182
Approved by: https://github.com/ezyang
Summary:
We've made the following changes:
- The new way to use the API is `m.impl_abstract_pystub(module, context)`.
Every subsequent m.def of an op inside the TORCH_LIBRARY block gives
the op the `impl_abstract_pystub`.
- Added a mechanism to determine if an operator was defined in Python or C++.
Library.define in Python appends the op to a global set, which is analogous
to what we do for tracking Library.impl.
- If someone does `torch.library.impl_abstract` in Python for an operator, then
we require that it has an `impl_abstract_pystub` specified and we also check
that the module in the `impl_abstract_pystub` is the same as the module where
the call to `torch.library.impl_abstract` exists.
- Unfortunately we can't check the "context" (which is the buck target on
buck-based systems) because buck sits above us.
Test Plan: - existing tests
Differential Revision: D50972148
Pull Request resolved: https://github.com/pytorch/pytorch/pull/112851
Approved by: https://github.com/ezyang
This should be the last of the "it used to work with static shapes but
it doesn't work with dynamic shapes" hard errors. Now we will just
specialize if you hit it from C++.
The strategy here is a bit clever. We shunt the size() call to Python
binding if an error would have occurred. Importantly, we already have
logic to make sure the newly allocated ints stay live for the duration
of the ArrayRef access.
storage_offset is intentionally omitted because there are some problems
with it. I will fix them next.
This should let us get rid of the aotautograd_static test configuration.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111935
Approved by: https://github.com/zou3519
torch.library.impl now accepts a device string (e.g. "cpu", "cuda"). It
still accepts DispatchKey strings, but we no longer document this, because
using arbitrary DispatchKeys is more for the power users.
We map the device string to a DispatchKey and then register the impl for
said DispatchKey. A user may also specify multiple device strings at once
or specify "types=default" to get a CompositeExplicitAutograd registration.
Test Plan:
- new tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111659
Approved by: https://github.com/soulitzer
ghstack dependencies: #111380
This PR supports sym_ite. This is useful for converting SymBool to SymInt in e.g. #109916. Internally, it uses sympy.Piecewise. We cannot use sympy.ITE because it expects the arguments and output all to be boolean type but we want return SymInt type when converting a SymBool to SymInt. So we use sympy.Piecewise to denote the symbolic relationship.
Note that this pr uses the range analysis for sympy.Piecewise implemented in https://github.com/pytorch/pytorch/blob/main/torch/utils/_sympy/value_ranges.py.
Test Plan:
See added test.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111440
Approved by: https://github.com/ezyang
Fixes the string_view errors and reland the work. The previous changes in torch/csrc/utils/invalid_arguments.cpp were too aggressive and not tested thoroughly. They are discarded.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/110518
Approved by: https://github.com/ezyang
We want to be able to use SingletonSymNode to represent strides for Jagged layout tensor. The following is for 3D, but easily generalizable to higher dimensions.
Constraints:
- [B, x, D] (where x represents the "variably lengthed dim") can be strided in two ways [x, 1, sum(x)] and [dx, d, 1]. We need two different placeholder values depending on how the jagged tensor is strided.
- When doing operations we need the strides of output tensors to be expressable in terms of the strides and sizes of the inner tensors. Given [B, x, D] @ [D, D'], the output strides is [x * D', D', 1] rather than some opaque [x2, D', 1]. This constraint exists because if I'm tracing, I need a symint to represent the output stride. This symint needs to come from somewhere; I get it in several ways: (1) create a constant, (2) unbacked symint, (3) create a new input using a source, (4) output of an operation on an existing symint. It is clear that (4) is what we want here, which brings us to the design below.
Design:
Given the two constraints, the most straightforward way to implement this is actually to update SingletonSymNode to include some scalar factor, i.e. Morally, SingletonSymNode represents `factor * [s_0, s_1, …, s_n]` This enables us to symbolically compute strides from sizes.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/110369
Approved by: https://github.com/ezyang
ghstack dependencies: #110044
This PR enables the misc-XX checks in clang-tidy. Meanwhile, I excluded some of them that require a lot of code changes and have no immediate benefits. Some additional fixes and suppression were also given.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/110283
Approved by: https://github.com/albanD
I added some tests for Conj, Neg and ZeroTensor for both python and C++ functionalization. This also fixes a nasty segfult when running a functorch `jacfwd` test with `torch.compile`, once AOTAutograd is using `FunctionalTensor`.
Changes:
(1) I use Jeffrey's `make_wrapper_subclass(extra_dispatch_keys)` kwarg to plumb extra dispatch keys ontoto the wrapper, mirroring what C++ functionalization does (C++ functionalization will mirror all dispatch keys from the inner tensor to the wrapper, except for python and functorch keys).
(2) FunctionalTensorMode will decompose CompositeImplicitAutograd ops, since (for example) ZeroTensor kernels can send ops like `.to()` directly to the Python key. We'll need a way to toggle this later for pre-dispatch functionalization
(3) Bound `_ForceDispatchKeyGuard` and BatchedTensorImpl's dispatch keyset to python
Pull Request resolved: https://github.com/pytorch/pytorch/pull/109023
Approved by: https://github.com/zou3519
ghstack dependencies: #108654, #109662, #109632
In this PR:
- When Constant SymNode are detected in unary/binary ops demote them to plain int/bool before proceeding. Sometimes this means doing a unary op with a Constant SymNode would result in a plain bool.
- Introduce an is_symbolic method, only available from Python. We need this because isinstance(x, SymInt) is no longer sufficient to check whether a given int/SymInt is symbolic or not. See later PR in the stack to see how this is used.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/109169
Approved by: https://github.com/ezyang
This PR adds a new `FunctionalTensor` subclass, and `FunctionalTensorMode` torch dispatch mode. Together, this class/mode are a lightweight wrapper around our existing C++ functionalization logic.
This idea came from Ed - later in the stack, I want to be able to run functionalization **underneath** torch_dispatch, when performing tracing in AOTAutograd. I can't do this easily with vanilla C++ functionalization, because it has a dedicated dispatch key that always runs before TorchDispatch. However, by adding a torch_dispatch mode shim around functionalization, we can use functionalization as a torch_dispatch mode, which will make it easier to run underneath other modes later.
This PR provides the basic new classes, and some light testing.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/106404
Approved by: https://github.com/ezyang
