mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-07 12:21:27 +01:00
3ef031909f
### Background:
`set(x,y)` changes the untyped storage of x to be the same as y.
```python
import torch
from torch._subclasses.fake_tensor import FakeTensorMode
x1 = torch.ones(2,3)
y1 = torch.ones(2,3)
z1 = torch.ops.aten.set_.source_Tensor(x1, y1)
fake_tensor_mode = FakeTensorMode()
x2 = fake_tensor_mode.from_tensor(torch.ones(2,3))
y2 = fake_tensor_mode.from_tensor(torch.ones(2,3))
z2 = torch.ops.aten.set_.source_Tensor(x2, y2)
print(f"x1: {x1.untyped_storage()._cdata}, y1: {y1.untyped_storage()._cdata}, z1: {z1.untyped_storage()._cdata}")
print(f"x2: {x2.untyped_storage()._cdata}, y2: {y2.untyped_storage()._cdata}, z2: {z2.untyped_storage()._cdata}")
# x1: 99973024, y1: 99973024, z1: 99973024
# x2: 112107232, y2: 112107232, z2: 112107232
```
### Error before this diff
Consider this example:
```python
import torch
def fn(x):
p = torch.nn.Parameter(x + 123)
return p, p.sin()
opt = torch.compile(fn, fullgraph=True)
x = torch.ones(16, device="cuda", requires_grad=True)
p, r = opt(x)
r.sum().backward()
```
When running with `TORCH_LOGS=aot`, we have `set_` in the graph.
```
def forward(self, primals_1: "f32[16][1]cuda:0", primals_2: "f32[16][1]cuda:0"):
# File: /home/boyuan/playground/inductor/donated_buffer.py:4 in fn, code: p = torch.nn.Parameter(x + 123)
add: "f32[16][1]cuda:0" = torch.ops.aten.add.Tensor(primals_1, 123); primals_1 = None
# File: /home/boyuan/playground/inductor/donated_buffer.py:5 in fn, code: return p, p.sin()
sin: "f32[16][1]cuda:0" = torch.ops.aten.sin.default(add)
# No stacktrace found for following nodes
set_: "f32[16][1]cuda:0" = torch.ops.aten.set_.source_Tensor(primals_2, add); primals_2 = set_ = None
return (sin, add)
```
`set_: "f32[16][1]cuda:0" = torch.ops.aten.set_.source_Tensor(primals_2, add)` should change the storage of `primals_2` to be the same as `add`. However, this is not true before this diff. We found different untyped_storage() for meta['val'] of `set_`, `add`, and `primals_2`.
This also leads to an error with donated buffer (#130580), which checks alias by untyped_storage. Since `add` and `primals_2` have different untyped_storage (which is wrong), add is wrongly marked as donated buffer.
### Root Cause
During tracing, we have args, kwargs, out, and proxy_args, proxy_kwargs, proxy_out.
We use args and kwargs to compute `out = func(*args, **kwargs)` ([Here](https://github.com/pytorch/pytorch/blob/main/torch/fx/experimental/proxy_tensor.py#L912)). Later, we set out to its proxy, essentially calling `proxy_out.node.meta["val"] = out.detach()`.
Due to the detach, the storage change happens on args but not on proxy_args.node.meta["val"] when func is torch.ops.aten.set_. I repro'ed this behavior of detach in eager code.
```python
import torch
x = torch.ones(2,3)
x_detach = x.detach()
y = torch.ones(2,3)
z = torch.ops.aten.set_.source_Tensor(x_detach, y)
print(f"x: {x.untyped_storage()._cdata}, x_detach: {x_detach.untyped_storage()._cdata}, y: {y.untyped_storage()._cdata}, z: {z.untyped_storage()._cdata}")
# x: 97023632, x_detach: 97026480, y: 97026480, z: 97026480
```
To fix the issue, this PR manually resets node.meta["val"] if the storage has changed.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/141308
Approved by: https://github.com/bdhirsh
|
||
|---|---|---|
| .. | ||
| backends | ||
| polyfills | ||
| repro | ||
| variables | ||
| __init__.py | ||
| _trace_wrapped_higher_order_op.py | ||
| bytecode_analysis.py | ||
| bytecode_transformation.py | ||
| cache_size.py | ||
| callback.py | ||
| code_context.py | ||
| codegen.py | ||
| compiled_autograd.py | ||
| comptime.py | ||
| config.py | ||
| convert_frame.py | ||
| create_parameter_op.py | ||
| current_scope_id.py | ||
| debug_utils.py | ||
| decorators.py | ||
| device_interface.py | ||
| distributed.py | ||
| eval_frame.py | ||
| exc.py | ||
| external_utils.py | ||
| funcname_cache.py | ||
| guards.py | ||
| hooks.py | ||
| logging.py | ||
| metrics_context.py | ||
| mutation_guard.py | ||
| output_graph.py | ||
| pgo.py | ||
| profiler.py | ||
| replay_record.py | ||
| resume_execution.py | ||
| side_effects.py | ||
| source.py | ||
| symbolic_convert.py | ||
| tensor_version_op.py | ||
| test_case.py | ||
| test_minifier_common.py | ||
| testing.py | ||
| trace_rules.py | ||
| types.py | ||
| utils.py | ||