Add some on-exit logging to the async compile workers. When you use `TORCH_LOGS=async_compile` (or `all`) it will now report how many workers were enqueued & dequeued (should be the same) as well as queuing time (how long workers sat on the queue before starting to run) and maximum depth (how many workers were waiting to start.
Tested manually by running a larger internal model and then lowering the number of available workers to see the time and depth get longer.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/155820
Approved by: https://github.com/masnesral
In order to take the globally best tiling, we need to normalize all the node read and writes to a common iteration space. This first pr finds a common split among nodes in a fused scheduler node, and then normalizes reads and writes to the common split.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153723
Approved by: https://github.com/jansel
https://github.com/pytorch/pytorch/pull/152708 expanded support of `get_estimated_runtime` to many more types of `SchedulerNodes`. This caused an increase in compile time because we're always calling `get_estimated_runtime` to populate the metrics table. This PR adds a flag for this logging, which reduces the instruction count by 8%. Long term, we should probably merge metrics.py with TORCH_LOGS/tlparse (suggestion from @xmfan).
Update: added support for TORCH_LOGS for the metrics logging.
Test Plan:
mm_loop.py and many existing tests cover.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153506
Approved by: https://github.com/eellison
Fixes#153646
This PR refactors the logging behavior in the FX pass insert_deferred_runtime_asserts and runtime_assert.py to separate verbose/intermediate graph logs from the final output graph log. All verbose logs generated during the FX pass are now routed to a new artifact logger, graph_code_verbose, while only the final output graph remains logged to the original graph_code artifact.
Changes
- Added a new artifact logger: [graph_code_log = torch._logging.getArtifactLogger(__name__, "graph_code_verbose")]
- Updated all verbose/intermediate FX pass logs in [insert_deferred_runtime_asserts] to use the new graph_code_verbose artifact.
- Ensured that only the final output graph is logged to the original graph_code artifact.
- No changes to the FX pass logic or output—only logging behavior is affected.
Notes
This change is backward-compatible and does not affect the functional behavior of FX passes.
No changes to user-facing APIs.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153775
Approved by: https://github.com/williamwen42
Summary:
Add new structured logging "inductor_pre_grad_graph"
This is for inductor provenance tracking front-end to load this graph from tlparse.
ghstack-source-id: 257581974
exported-using-ghexport
Test Plan:
```
buck2 run 'fbcode//mode/dev-nosan' //caffe2/test/dynamo:test_dynamo -- -r StructuredTraceTest
```
Differential Revision: D67150288
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143126
Approved by: https://github.com/desertfire
Summary:
- Added TORCH_LOGS=cache to dump cache stats on exit - supported by RemoteCache.
- Split REMOTE_CACHE_VERSION - it was used for both JKs fx_graph_memcache_version and autotune_memcache_version but they really should be separate (just in case we need to change one but not the other)
- Prepare `_ManifoldCache` for use with other subpath keys
- Move create_cache to be more public and use it in codecache
- Add _InductorMetaTy alias (still just a dict)
- Cleaned up some common cached_autotune calls in triton_heuristics
Test Plan: unit tests
Reviewed By: oulgen
Differential Revision: D62648249
Pull Request resolved: https://github.com/pytorch/pytorch/pull/136456
Approved by: https://github.com/oulgen
Summary:
A bit of refactoring to prepare to remove `None` as a way to specify static dimensions in dynamic shapes, given we already have `Dim.STATIC` for the same purpose. We will now warn whenever this happens. However no tests were modified because problematic uses of `None` still need to behave as they do today, until we are ready to remove support. It should be easy to port tests by replacing the warning function to raise instead.
Note that other uses of `None`, such as for entire values (tensor or non-tensor) remain as is. Moving forward this should be the only purpose of `None` (at least externally).
Finally, there's a bit of confusion in our representation now because `AUTO` also internally transforms to `None`. Renamed dynamic_shapes to transformed_dynamic_shapes where this happens. Overall the two forms (pre and post transformation) have different properties so should probably not be represented in the same format in the future.
Test Plan: existing
Differential Revision: D62040729
Pull Request resolved: https://github.com/pytorch/pytorch/pull/134877
Approved by: https://github.com/pianpwk
Restart the work from PR https://github.com/pytorch/pytorch/pull/100331 in this new PR since it's hard to rebase. It would be expected that some code is copy/pasted from the previous PR and main idea is the same.
Previously we see relatively large compilation time increase due to too many loop orders being considered. This PR tries to continue the work by doing pruning and only considering loop orders that we know for sure are relevant (i.e. do it on demand).
Some manually created cases that loop ordering matters are added as unit tests. The PR can make sure inductor does not miss fusion opportunities for them.
This PR should solve the not-able to fusion problem in https://github.com/pytorch/pytorch/issues/130015
Right now there is still significant increase of compilation time. I'll disable the feature by default. Later on after the compilation time issue is resolved, I'll enable it by default.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/126254
Approved by: https://github.com/jansel
Summary:
Today there is no good mechanism to detect progress of non-strict export line-by-line in user code. This caused some pain recently in trying to find the exact line of user code that was triggering a bug where the process appeared stuck because deep down something was calling some symbolic shapes code that was suffering some exponential blowup.
This PR adds a environment variable for extended debugging that will log the line of user code corresponding to every torch function call. It only works in non-strict export for now. Prefix setting this environment variable with `TORCH_LOGS` enabled for `export` logs at `DEBUG` level (i.e., with a `+` prefix), i.e.,.:
```
TORCHEXPORT_EXTENDED_DEBUG_CURRENT_LOC=1 TORCH_LOGS="+export" ...
```
This will show logs with something like:
```
...
prim::device called at .../example.py:4284 in foo
TensorBase.item called at .../example.py:4277 in bar
...
```
We already have an existing place to intercept torch functions where we process data-dependent errors in non-strict, so parking the logging there. An alternative place we could be doing this is where we add `stack_trace` metadata when generating code, but unfortunately at least the example that motivated this gets stuck before generating code, so that would be too late.
Test Plan: ran it on some sample commands
Differential Revision: D61692156
Pull Request resolved: https://github.com/pytorch/pytorch/pull/134298
Approved by: https://github.com/angelayi
Support of effectful operations in backward:
1/ AOTD collects metadata from forward fn only, so we can have usage of effectful ops in backward, that were not used in forward => Allowing tokens discovery during joint function .
FunctionalTensorMode holds _tokens, in Joint function after tracing forward we memoize _tokens as `_tokens_forward_output`.
2/ Tokens are added as primals inputs (forward) in EffectTokensWrapper.
Tokens that will be used in backward are in partitioner saved values. We do not have control on which positions they are saved in forward outputs.
2/ If new tokens discovered in backward after tracing joint_fn, the result graph will be manually added in the end of primals.
_aot_autograd/utils.py
3/ All effectful ops during backward are marked with 'must_be_in_backward' partitioner_tag, to prevent partiitoner to place them in forward.
For that functional_tensor_mode got new optional state `self._effects_partitioner_tag` for effectful ops, to set after tracing forward.
There are additional changes in partitioner to improve functionality of 'must_be_in_backward'
4/ Unlift tokens now should run for both forward and backward.
- As saved for backward tokens are placed on non static places - we identify input and output tokens to erase, by input and output of `with_effects` operation
- In forward we can have input tokens, discovered in backward, that are not used in with_effects ops in forward, but saved for backward. We identify them by position in forward inputs.
5/ Adding aot debug logging for graphs before unlifting and before adding additional primal for backward tokens.
Tests:
```
python test/higher_order_ops/test_with_effects.py
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132638
Approved by: https://github.com/bdhirsh
move benchmarking out of `torch._inductor.runtime.runtime_utils` and into `torch._inductor.runtime.benchmarking`, and prefer this path over directly accessing Triton's benchmarking
Fixes #ISSUE_NUMBER
Pull Request resolved: https://github.com/pytorch/pytorch/pull/132827
Approved by: https://github.com/eellison
Adds support for SymInts in the FakeTensor cache.
A couple notes:
1. When a SymInt is present in the input key for a FakeTensor operation we cache on the ShapeEnv instead of using the FakeTensorMode cache. This is necessary so we don't have to remember and check the guards. It reduces the cache hits but there's diminishing return on how much work we can do before the cache becomes more of a burden than a gain.
2. We need to be careful that when we cache an output SymInt that is a direct copy from the input that when we have a cache-hit we copy the SymNode from the input to the output. This is important because the fx-graph building code actually uses SymNode ids in the process of building the graph so constructing a same-content-but-different-id SymNode will fail.
3. In the cache key we store SymInts as a _PySymInputStub. These represent SymInt (and friends) but support `__hash__` and `__eq__` (which SymInt do not).
4. In the cache entry we store SymInts as a _SymIntOutputStub.
Perf example:
```
python benchmarks/dynamo/timm_models.py --ci --accuracy --timing
--explain --inductor --dynamic-shapes --dynamic-batch-only --device cuda
--training --amp --total-partitions 2 --partition-id 0 --output
/tmp/training_timm_models.csv --filter crossvit_9_240
```
fake tensor cache before:
```
INFO: FakeTensor cache stats:
INFO: cache_hits: 68137
INFO: cache_misses: 837
INFO: cache_bypasses:
INFO: symbolic shape: 48224
INFO: CompositeImplicitAutograd: 917
INFO: non-fake tensor: 70
INFO: non-FakeTensor output: 62
INFO: non-builtin: 8
INFO: dynamic output shape: 1
```
and after:
```
INFO: FakeTensor cache stats:
INFO: cache_hits: 88187
INFO: cache_misses: 14233
INFO: cache_bypasses:
INFO: CompositeImplicitAutograd: 1037
INFO: non-FakeTensor output: 602
INFO: non-fake tensor: 70
INFO: unsafe view: 36
INFO: non-builtin: 8
INFO: dynamic output shape: 1
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/127596
Approved by: https://github.com/eellison
ghstack dependencies: #131014, #129780
Summary:
Add '`TORCH_LOGS=+fsdp`' in the CLI to print fsdp logs
Example:
`TORCH_LOGS=+fsdp torchrun --standalone --nproc_per_node=2 run_fsdp.py`
Description:
Add logging to `FSDPParamGroup.pre_forward`, `FSDPParamGroup.post_forward`, `FSDPParamGroup.pre_backward`, and `FSDPParamGroup.post_backward`, `FSDPState._root_pre_forward` if is the root, and `FSDPState._root_post_backward_final_callback`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/128663
Approved by: https://github.com/weifengpy, https://github.com/awgu
e.g. dist_ddp -> ddp
'distributed' shortcut remains unchained
Feedback has been that it is not appealing to have the dist_ prefix,
and the main reason for it was to keep the distributed shortcuts grouped
together in the help menu. It's nice to have shorter shortcuts.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/126499
Approved by: https://github.com/XilunWu, https://github.com/kwen2501
ghstack dependencies: #126322
- sets it as a fake stack trace as we don't have a generic comment feature
- when verbose is disabled, still adds a contextmanager and flag checks. the alternative is to use MACROS, but that wouldn't be usable with TORCH_LOGS
Pull Request resolved: https://github.com/pytorch/pytorch/pull/124954
Approved by: https://github.com/jansel
Partially fixes https://github.com/pytorch/pytorch/issues/105077
Repro:
```python
import tempfile
import torch
from torch._subclasses import fake_tensor
class TheModelClass(torch.nn.Module):
def __init__(self):
super(TheModelClass, self).__init__()
self.fc1 = torch.nn.Linear(5, 10)
def forward(self, x):
return self.fc1(x)
with tempfile.NamedTemporaryFile() as state_dict_file:
# Create state_dict to be loaded later
model = TheModelClass()
torch.save(model.state_dict(), state_dict_file.name)
fake_mode = fake_tensor.FakeTensorMode()
with fake_mode:
# This is where the bug is triggered
state_dict = torch.load(state_dict_file.name)
```
Error:
```bash
Traceback (most recent call last):
File "issue_gh_torch_105077.py", line 22, in <module>
state_dict = torch.load(state_dict_file.name)
File "/opt/pytorch/torch/serialization.py", line 1014, in load
return _load(opened_zipfile,
File "/opt/pytorch/torch/serialization.py", line 1422, in _load
result = unpickler.load()
File "/opt/pytorch/torch/_utils.py", line 205, in _rebuild_tensor_v2
tensor = _rebuild_tensor(storage, storage_offset, size, stride)
File "/opt/pytorch/torch/_utils.py", line 184, in _rebuild_tensor
return t.set_(storage._untyped_storage, storage_offset, size, stride)
File "/opt/pytorch/torch/utils/_stats.py", line 20, in wrapper
return fn(*args, **kwargs)
File "/opt/pytorch/torch/_subclasses/fake_tensor.py", line 1288, in __torch_dispatch__
return self.dispatch(func, types, args, kwargs)
File "/opt/pytorch/torch/_subclasses/fake_tensor.py", line 1468, in dispatch
self.invalidate_written_to_constants(func, flat_arg_fake_tensors, args, kwargs)
File "/opt/pytorch/torch/_subclasses/fake_tensor.py", line 1733, in invalidate_written_to_constants
_, new_kwargs = normalize_function(
File "/opt/pytorch/torch/fx/operator_schemas.py", line 297, in normalize_function
torch_op_schemas = get_signature_for_torch_op(target)
File "/opt/pytorch/torch/fx/operator_schemas.py", line 167, in get_signature_for_torch_op
signatures = [_torchscript_schema_to_signature(schema) for schema in schemas]
File "/opt/pytorch/torch/fx/operator_schemas.py", line 167, in <listcomp>
signatures = [_torchscript_schema_to_signature(schema) for schema in schemas]
File "/opt/pytorch/torch/fx/operator_schemas.py", line 70, in _torchscript_schema_to_signature
arg_type = _torchscript_type_to_python_type(arg.type)
File "/opt/pytorch/torch/fx/operator_schemas.py", line 64, in _torchscript_type_to_python_type
return eval(ts_type.annotation_str, _type_eval_globals)
File "<string>", line 1, in <module>
NameError: name 'Storage' is not defined
```
This PR adds the ability to create fake tensors during `torch.load` by wrapping the `torch.tensor.set_` call around a `torch.utils._mode_utils.no_dispatch()` to skip fake mode dispatcher for it and thus create a real tensor. It later calls `fake_mode.from_tensor(t)` to finally create the fake tensor.
Co-authored-by: Edward Z. Yang <ezyang@mit.edu>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/108186
Approved by: https://github.com/ezyang
This PR is the start to enable the integrate pytorch distributed logs in Torch LOGs. We now already have one tag "distributed" for all distributed components but distributed is a very large component and we want to have some hierarchy and give users options to only turn on logs for certain submodules. So we also added tags starting with "dist_*" for each submodule. (This PR only adds some of them and we are going to add more down the road)
Related discussions can be found here: https://github.com/pytorch/pytorch/issues/113544
Pull Request resolved: https://github.com/pytorch/pytorch/pull/116434
Approved by: https://github.com/awgu, https://github.com/wanchaol
There are now 3 ways to see logs from ddpoptimzer.
1) TORCH_LOGS="distributed"
2) TORCH_LOGS="dynamo"
3) TORCH_LOGS="torch._dynamo.backends.distributed"
(1 and 2 are different supersets of 3 that also include other content)
Note: ddp_graphs is still a separate 'artifact' logger, which just
includes graph dumps from the graph-splitting process.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/114376
Approved by: https://github.com/wanchaol
Followup to https://github.com/pytorch/pytorch/pull/110325 - re-add the `report_all_guard_failures config` as a logging artifact `recompiles_verbose` with the following changes:
- evaluating the check must be wrapped with exception handling because subsequent code parts following the first failure may result in errors if evaluated (e.g. if a guard checks first for size, then tries to index - a guard failure due to insufficient size would result in an index error for the latter check).
- Adding a test for this case
Sample:
```python
import torch
def fn(x):
return torch.rand(x[-1], len(x))
opt_fn = torch.compile(fn)
opt_fn([4, 5, 6])
opt_fn([7, 8])
opt_fn([9])
```
Output (with `TORCH_LOGS="recompiles_verbose"`):
```bash
[2023-11-15 16:13:26,741] torch._dynamo.guards.__recompiles_verbose: [DEBUG] Recompiling function fn in /data/users/williamwen/pytorch/playground5.py:15
[2023-11-15 16:13:26,741] torch._dynamo.guards.__recompiles_verbose: [DEBUG] triggered by the following guard failure(s):
[2023-11-15 16:13:26,741] torch._dynamo.guards.__recompiles_verbose: [DEBUG] guard 0 failures:
[2023-11-15 16:13:26,741] torch._dynamo.guards.__recompiles_verbose: [DEBUG] - len(L['x']) == 3
[2023-11-15 16:13:26,741] torch._dynamo.guards.__recompiles_verbose: [DEBUG] - L['x'][0] == 4
[2023-11-15 16:13:26,741] torch._dynamo.guards.__recompiles_verbose: [DEBUG] - L['x'][1] == 5
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] Recompiling function fn in /data/users/williamwen/pytorch/playground5.py:15
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] triggered by the following guard failure(s):
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] guard 0 failures:
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] - len(L['x']) == 2
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG]
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] guard 1 failures:
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] - len(L['x']) == 3
[2023-11-15 16:13:26,970] torch._dynamo.guards.__recompiles_verbose: [DEBUG] - L['x'][0] == 4
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113585
Approved by: https://github.com/jon-chuang, https://github.com/ezyang
When SymNode was refactored into its own module, this broke logging for this file, as the `dynamic` alias no longer covered it. This PR adds supports for an alias to point to multiple qualified module names. To drive the refactor, I renamed `log_alias_to_log_qname` to `log_alias_to_log_qnames` and then audited all use sites. I invite you to do so as well.
For good measure, I also add dynamic to dynamo, so that I always get dynamic logs when dynamo is enabled. Empirically this will be helpful because people keep sending me dynamo debug logs that don't have dynamic logs.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/113567
Approved by: https://github.com/Skylion007, https://github.com/lezcano, https://github.com/mlazos
ghstack dependencies: #113566
