pytorch/test/dynamo
Shunting Zhang 4cc64d6234 [inductor] pre grad graph bisecting (#166344)
A few things to note:
1. Customers like vllm use a custom backend (e.g. VllmBackend), split the graph, and call standalone_compile for each split. If we let the bisector override the backend, we won't bisect thru the custom backend. `test_configs.bisect_keep_custom_backend_for_inductor` is used to keep the custom backend if we are bisecting for inductor.
2. pre_grad_graph bisecting and lowering bisecting so far does not compose well with each other since an issue may be just captured by the first one we try. `test_configs.bisect_pre_grad_graph` is used to enable the 'pre_grad_graph' bisecting.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166344
Approved by: https://github.com/eellison
2025-11-01 09:22:21 +00:00
..
cpython/3_13 [2/N][Fix] Fix typo in test folder (#166374) 2025-10-29 03:02:07 +00:00
mock_modules
__init__.py
_test_nested_graph_breaks_helper.py [dynamo, nested graph breaks] add nested graph break tests (#144516) 2025-08-27 03:00:56 +00:00
imports_non_circular_repro.py [Bugfix] Fix circular import between export and dynamo from tensor fn map (#158931) 2025-07-24 22:24:57 +00:00
mock_store_global_crossfile_inline.py
test_activation_checkpointing.py [dynamo, nested graph breaks] disallow nested graph breaks in HOPs (#166016) 2025-10-28 03:03:38 +00:00
test_after_aot.py
test_aot_autograd_cache.py Add regional aot eager support to AOTAutogradCacheEntry (#166650) 2025-10-31 18:54:09 +00:00
test_aot_autograd.py Revert "Fix comparing inductor actual strides vs bw graph for activations should not throw DDE. (#166277)" 2025-10-30 15:49:23 +00:00
test_aot_compile.py [precompile] Preserve default arguments for dynamo capture (#166654) 2025-11-01 00:12:10 +00:00
test_autograd_function.py Enable XPU for test_autograd_function.py (#160309) 2025-08-13 06:38:34 +00:00
test_backends.py [Code Clean] Remove deadcodes about Python3.9 [2/N] (#163627) 2025-09-24 07:30:50 +00:00
test_backward_higher_order_ops.py [reland] Allow setting grad_dtype on leaf tensors (#164751) 2025-10-08 20:23:13 +00:00
test_base_hop.py Fix requires_cuda to requires_cuda_and_triton (#160222) 2025-08-10 07:05:52 +00:00
test_base_output.py
test_buffers_override.py Typo fixes for "overridden" in comments and function names (#155944) 2025-06-14 03:37:38 +00:00
test_bytecode_utils.py [dynamo, 3.14] fix misc. bugs to get most dynamo unittests passing locally in 3.14 (#164631) 2025-10-28 03:24:22 +00:00
test_callback.py Use dataclass features in two classes (#164221) 2025-10-01 03:20:39 +00:00
test_compile.py [User-streams] Make torch.Event weakref compatible (#164522) 2025-10-29 23:06:31 +00:00
test_compiler_bisector.py [inductor] pre grad graph bisecting (#166344) 2025-11-01 09:22:21 +00:00
test_comptime.py [aotd] Support saved tensors hooks in aot_autograd (#150032) 2025-05-22 14:09:38 +00:00
test_config.py
test_ctx_manager.py Revert "[user-streams] Fix stream graph output semantics (#164819)" 2025-10-30 16:53:32 +00:00
test_cudagraphs_expandable_segments.py
test_cudagraphs.py
test_debug_utils.py [dynamo, 3.14] fix segfault due to improper create_call_function_ex (#166678) 2025-10-31 20:44:53 +00:00
test_decorators.py [dynamo] fix error_on_graph_break bug where non-empty checkpoint results in unwanted graph break resumption (#166586) 2025-10-31 03:36:27 +00:00
test_deque_reconstruct.py
test_deviceguard.py
test_dicts.py [4/N] Remove unused loop variables in tests (#166690) 2025-10-31 10:20:48 +00:00
test_dynamic_shapes.py
test_einops.py Add einops x torch.compile testing in PyTorch CI (#157416) 2025-07-03 17:36:39 +00:00
test_error_messages.py [dynamo] clarify graph break handling/logging in symbolic_convert (#166587) 2025-10-31 20:13:16 +00:00
test_exc.py [dynamo] update pytorch-labs -> meta-pytorch in graph break URLs (#159975) 2025-08-06 23:57:31 +00:00
test_exceptions.py [dynamo] Fix python_type for UserDefinedClassExceptionVariable (#166251) 2025-10-27 16:47:32 +00:00
test_export_mutations.py [export] Turn on install_free_tensors flag (#164691) 2025-10-14 15:33:50 +00:00
test_export.py Support python slicing with tensor inputs. (#165074) 2025-10-29 01:18:45 +00:00
test_fake_distributed.py [dynamo][device_mesh] Support mesh_dim_names (#164200) 2025-09-30 07:16:28 +00:00
test_flat_apply.py [export] Add meta[val] to getattr nodes (#154934) 2025-06-13 05:48:21 +00:00
test_frame_init.py Use dataclass features in two classes (#164221) 2025-10-01 03:20:39 +00:00
test_functions.py Revert "Fix torch.full with dynamic tensor fill_value in torch.compile (#166554)" 2025-10-31 12:55:31 +00:00
test_fx_annotate.py [dynamo][annotate] Graph break cleanly on fx.traceback.annotate reconstruction (#166006) 2025-10-21 20:48:04 +00:00
test_fx_graph_runnable.py [devx] Fix invalid symbol definition emitted in fx_graph_runnable.py (#166529) 2025-10-30 16:40:12 +00:00
test_fx_passes_pre_grad.py
test_generator.py [dynamo, nested graph breaks] disable nested graph breaks in generators; enable nested_graph_breaks in test_ctx_manager.py and test_generator.py (#165809) 2025-10-28 03:03:37 +00:00
test_global.py
test_graph_deduplication.py [Dynamo][Hierarchical Compile] Flatten tuple inputs for regions (#158812) 2025-08-18 18:03:11 +00:00
test_graph_region_tracker.py [ROCm] Remove HIPBLASLT_ALLOW_TF32 from codebase (#162998) 2025-09-18 13:53:48 +00:00
test_guard_manager.py [dynamo] Remove duplicated guards (#165806) 2025-10-20 05:50:33 +00:00
test_guard_serialization.py [dynamo][guards] CLASS_MATCH guard for readability (#166217) 2025-10-26 18:35:27 +00:00
test_higher_order_ops.py [4/N] Remove unused loop variables in tests (#166690) 2025-10-31 10:20:48 +00:00
test_hooks.py [4/N] Remove unused loop variables in tests (#166690) 2025-10-31 10:20:48 +00:00
test_inline_and_install.py [export] Turn on install_free_tensors flag (#164691) 2025-10-14 15:33:50 +00:00
test_input_attr_tracking.py [BE]: Update ruff to 0.11.8 (#153249) 2025-05-12 18:30:52 +00:00
test_install_free_tensors.py
test_interop.py [dynamo] Properly handle torch.script.jit under @staticmethod (#153984) 2025-05-21 19:45:06 +00:00
test_list.py [dynamo] Add __iter__ for iterable VariableTrackers (#166349) 2025-10-29 21:54:37 +00:00
test_logging.py [ROCm][CI] forward fix kineto submodule bump (#166421) 2025-10-28 17:40:23 +00:00
test_metrics_context.py [BE][5/6] fix typos in test/ (test/dynamo/) (#157639) 2025-07-06 06:34:25 +00:00
test_minifier.py Port two dynamo test cases for Intel GPU (#156056) 2025-06-19 12:49:04 +00:00
test_misc.py Revert "Make PT2 compile backprop through custom op without autograd key a hard error (#166367)" 2025-10-31 13:44:05 +00:00
test_model_output.py Port three dynamo test to Intel GPU (#156575) 2025-06-27 05:56:22 +00:00
test_modes.py [dynamo] Fix TorchFunctionMode handling with get_rng_state (#163412) 2025-09-23 15:37:45 +00:00
test_modules.py [3/N] Fix unused loop variables (#166509) 2025-10-30 20:13:51 +00:00
test_nested_graph_breaks.py [dynamo, nested graph breaks] add disable_nested_graph_breaks decorator/context manager (#166477) 2025-10-31 03:36:27 +00:00
test_nops.py
test_optimizers.py [BE][PYFMT] migrate PYFMT for test/[a-h]*/ to ruff format (#144555) 2025-06-24 04:53:54 +00:00
test_package.py [dynamo][guards] CLASS_MATCH guard for readability (#166217) 2025-10-26 18:35:27 +00:00
test_pgo.py [PGO] log missing sources in allowlist (#164881) 2025-10-09 04:39:09 +00:00
test_pre_dispatch.py
test_precompile_context.py Simplify PrecompileContext to no longer be a CacheArtifactManager (#162886) 2025-09-20 01:24:37 +00:00
test_profiler.py [dynamo][export] Do not graph break on torch.autograd._profiler_enabled for export (#164418) 2025-10-02 09:00:00 +00:00
test_python_autograd.py Enable all SIM rules except disabled ones (#164645) 2025-10-17 07:27:11 +00:00
test_python_dispatcher.py port 4 dynamo test files to Intel GPU (#157779) 2025-07-11 10:11:49 +00:00
test_recompile_ux.py More ruff SIM fixes (#164695) 2025-10-09 03:24:50 +00:00
test_recompiles.py Add logging for when inbuilt_inline_nn_modules will help with ID_MATCH guard triggered recompiles (#160592) 2025-08-15 17:09:39 +00:00
test_reconstruct.py [Linter] Expanding the scope of detecting device-bias code. (#159949) 2025-08-09 09:41:16 +00:00
test_regional_inductor.py Refactor AOTAutogradCacheEntry into AOTAutogradResult (#166656) 2025-10-31 18:54:09 +00:00
test_reorder_logs.py [dynamo] update pytorch-labs -> meta-pytorch in graph break URLs (#159975) 2025-08-06 23:57:31 +00:00
test_repros.py [2/N][Fix] Fix typo in test folder (#166374) 2025-10-29 03:02:07 +00:00
test_resume.py
test_sdpa.py SDP Backend function fix (#161169) 2025-09-19 20:19:59 +00:00
test_sets.py Capture TypeError in CONTAINS_OP (#161069) 2025-09-04 04:49:09 +00:00
test_skip_guard_eval_unsafe.py [dynamo][guards] Make class members go through obj.__class__.__dict__ (#159534) 2025-08-04 05:12:44 +00:00
test_skip_non_tensor.py
test_sources.py Enable ruff rule E721 (#165162) 2025-10-13 01:48:55 +00:00
test_streams.py [User-streams] Make torch.Event weakref compatible (#164522) 2025-10-29 23:06:31 +00:00
test_structured_trace.py Revert "Make PT2 compile backprop through custom op without autograd key a hard error (#166367)" 2025-10-31 13:44:05 +00:00
test_subclasses.py Update gm.print_readable to include Annotation (#165397) 2025-10-28 13:54:38 +00:00
test_subgraphs.py [BE][5/6] fix typos in test/ (test/dynamo/) (#157639) 2025-07-06 06:34:25 +00:00
test_torchrec.py
test_trace_rules.py Revert "[user-streams] Track symbolic current stream (#165212)" 2025-10-30 13:24:56 +00:00
test_unittest.py Add infra to run CPython tests under Dynamo (#150787) 2025-05-07 04:03:14 +00:00
test_unspec.py [4/N] Remove unused loop variables in tests (#166690) 2025-10-31 10:20:48 +00:00
test_utils.py [dynamo, 3.14] fix misc. bugs to get most dynamo unittests passing locally in 3.14 (#164631) 2025-10-28 03:24:22 +00:00
test_verify_correctness.py
test_view.py use check_size instead of check_is_size in ops.py (#164668) 2025-10-08 14:23:38 +00:00
utils.py [Dynamo] Introduce hook receiving list of traced code objects (#153622) 2025-05-28 15:40:09 +00:00