saymrwulf/pytorch
1
0
Fork 0
mirror of https://github.com/saymrwulf/pytorch.git synced 2026-05-14 20:57:59 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
pytorch/test/inductor
History
Rachel Guo da09984c0d [AOTI][Tooling][9/n] Add debug printer support for cpp kernel type (#136465) 
Summary:

As title.

Cpp kernel has a different codegen path: https://www.internalfb.com/code/fbsource/[6df946858879dd9bcefa18710dd79095a957f0dd]/fbcode/caffe2/torch/_inductor/codegen/cpp.py?lines=4643
Previously it is not wrapped/supported by the debug printer manager. This diff adds this support.
It can be useful for cpu models. See this for a use case: https://www.internalfb.com/phabricator/paste/view/P1598561051?lines=927

Test Plan:
```
AOT_INDUCTOR_DEBUG_INTERMEDIATE_VALUE_PRINTER=2 TORCHINDUCTOR_FORCE_DISABLE_CACHES=1  TORCHINDUCTOR_ABI_COMPATIBLE=1 TORCH_COMPILE_DEBUG=1 TORCH_LOGS="+graph, inductor, +schedule, output_code" buck2 run 'fbcode//mode/opt' fbcode//accelerators/workloads/models/slimdsnn:slimdsnn -- aot --batch-size 1
```

Differential Revision: D63053101

Pull Request resolved: https://github.com/pytorch/pytorch/pull/136465
Approved by: https://github.com/hl475
2024-09-26 17:30:43 +00:00
..
cpp
extension_backends [Inductor] Rename WrapperCodeGen to PythonWrapperCodegen (#136062) 2024-09-24 21:02:51 +00:00
__init__.py
CMakeLists.txt
custom_ops.cpp Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
indirect_assert_helper.py
minifier_smoke.py
mock_cache.py Add cache logs + other minor caching cleanup (#136456) 2024-09-24 14:00:23 +00:00
opinfo_harness.py
s429861_repro.py
test_aot_inductor.py [AOTI][Tooling][9/n] Add debug printer support for cpp kernel type (#136465) 2024-09-26 17:30:43 +00:00
test_aot_inductor_package.py [aoti] Add cpp loader (#135374) 2024-09-11 03:00:01 +00:00
test_aot_inductor_utils.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_auto_functionalize.py Remove onnx imports in dynamo (#136334) 2024-09-24 14:54:23 +00:00
test_autoheuristic.py
test_b2b_gemm.py
test_benchmark_fusion.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_benchmarking.py
test_binary_folding.py Fix autodeps from D62049222 that pyfmt broke (#136455) 2024-09-24 05:06:12 +00:00
test_ck_backend.py [ROCm][Inductor][CK] FP8 gemm (#136337) 2024-09-24 05:19:45 +00:00
test_codecache.py [Inductor UT] Generalize device-bias code introduced from #134874 and (#136596) 2024-09-26 02:56:59 +00:00
test_codegen_triton.py
test_combo_kernels.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_compile_worker.py
test_compiled_autograd.py Disable compiled autograd for re-entrant autograd (#135795) 2024-09-24 15:09:16 +00:00
test_compiled_optimizers.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_config.py
test_control_flow.py
test_coordinate_descent_tuner.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_cpp_wrapper_hipify.py [Inductor] Generalize cuda cpp wrapper as common triton based GPU cpp wrapper, will be reused by xpu in next PR. (#135312) 2024-09-11 23:59:54 +00:00
test_cpu_cpp_wrapper.py [AOTI] Support MKL linear ops in cpp wrapper (#134974) 2024-09-25 03:53:11 +00:00
test_cpu_repro.py [Inductor][CPP] Fix int8 cvt half (#136353) 2024-09-25 02:23:43 +00:00
test_cpu_select_algorithm.py [inductor] [cpp] fix max-autotune for single-thread dynamic shapes (#136418) 2024-09-25 09:24:05 +00:00
test_cuda_cpp_wrapper.py [AOTI] Support MKL linear ops in cpp wrapper (#134974) 2024-09-25 03:53:11 +00:00
test_cuda_repro.py Add Triton CPU as an Inductor backend (#133408) 2024-09-26 15:35:26 +00:00
test_cudacodecache.py
test_cudagraph_trees.py
test_cudagraph_trees_expandable_segments.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_custom_lowering.py
test_custom_post_grad_passes.py
test_cutlass_backend.py
test_debug_trace.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_decompose_mem_bound_mm.py
test_dependencies.py
test_distributed_patterns.py [Traceable FSDP2] Use .copy_ instead of .set_ for unsharded_param inplace update; Replace unsharded_param graph input usage with graph intermediate; Support FSDP2+LoRA (#133730) 2024-09-11 23:01:05 +00:00
test_efficient_conv_bn_eval.py Fix autodeps from D62049222 that pyfmt broke (#136455) 2024-09-24 05:06:12 +00:00
test_extension_backend.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_flex_attention.py [FlexAttention] Reduce expensive test time by 10x (#136677) 2024-09-26 16:40:21 +00:00
test_flex_decoding.py [ROCm] Enable Flex attention tests on AMD gpus (#136245) 2024-09-19 18:02:41 +00:00
test_foreach.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_fp8.py
test_fused_attention.py
test_fx_fusion.py
test_graph_transform_observer.py
test_group_batch_fusion.py
test_halide.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_indexing.py
test_inductor_freezing.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_inductor_utils.py
test_inplacing_pass.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_kernel_benchmark.py
test_layout_optim.py
test_loop_ordering.py [ROCm] skip test_fp8_cast_and_t on non-MI300 machines (#135917) 2024-09-13 02:46:48 +00:00
test_max_autotune.py update cache tests (#136215) 2024-09-21 20:36:22 +00:00
test_memory.py [Inductor UT] Generalize device-bias code introduced from #134874 and (#136596) 2024-09-26 02:56:59 +00:00
test_memory_planning.py Add Triton CPU as an Inductor backend (#133408) 2024-09-26 15:35:26 +00:00
test_metrics.py [pytorch][PR] [inductor] More fixes on the keys of constants and signature dictionaries (#136514) 2024-09-25 09:34:14 +00:00
test_minifier.py
test_minifier_isolate.py [Inductor UT] Generalize inductor UT for intel GPU (Part 2) (#134556) 2024-09-13 05:16:28 +00:00
test_mkldnn_pattern_matcher.py [Quant][PT2E]change flatten recipe for X86InductorQuantizer (#136298) 2024-09-24 04:30:12 +00:00
test_mmdecomp.py
test_move_constructors_to_cuda.py
test_multi_kernel.py
test_ordered_set.py
test_pad_mm.py Revert "Reland D62220158 (#136213)" 2024-09-19 12:44:54 +00:00
test_padding.py
test_pattern_matcher.py [Easy] Dont match to mm_plus_mm if not in max autotune (#135929) 2024-09-13 23:38:02 +00:00
test_perf.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_profiler.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_scatter_optimization.py
test_select_algorithm.py
test_smoke.py
test_snode_runtime.py
test_split_cat_fx_passes.py Don't run reshape pattern match on dynamic shape size tensor (#136100) 2024-09-17 15:08:55 +00:00
test_standalone_compile.py
test_torchbind.py
test_torchinductor.py [Inductor] Fix test_profiler_mark_wrapper_call_cuda_cuda_wrapper (#136356) 2024-09-20 18:35:09 +00:00
test_torchinductor_codegen_dynamic_shapes.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00
test_torchinductor_dynamic_shapes.py [Inductor] Rename WrapperCodeGen to PythonWrapperCodegen (#136062) 2024-09-24 21:02:51 +00:00
test_torchinductor_opinfo.py [Inductor UT] Generalize device-bias code introduced from #134874 and (#136596) 2024-09-26 02:56:59 +00:00
test_torchinductor_strided_blocks.py Increase TRITON_MAX_BLOCK['X'] (#135181) 2024-09-10 05:54:37 +00:00
test_triton_cpu_backend.py Add Triton CPU as an Inductor backend (#133408) 2024-09-26 15:35:26 +00:00
test_triton_extension_backend.py Fix autodeps from D62049222 that pyfmt broke (#136455) 2024-09-24 05:06:12 +00:00
test_triton_heuristics.py [inductor] ELEMENTS_PER_WARP_32 -> ONE_ELEMENT_PER_THREAD (#136472) 2024-09-25 18:21:09 +00:00
test_triton_kernels.py [Inductor UT] Generalize device-bias code introduced from #134874 and (#136596) 2024-09-26 02:56:59 +00:00
test_triton_wrapper.py
test_unbacked_symints.py
test_utils.py
test_xpu_basic.py Fix lint errors in fbcode (#135614) 2024-09-13 02:04:34 +00:00