[upstream #3539] [Bug Skip] test_cache_hot_load device=xpu fails on XPU: AOTAutograd / inductor cache counter mismatch#20
Draft
Stonepia wants to merge 1 commit into
Draft
[upstream #3539] [Bug Skip] test_cache_hot_load device=xpu fails on XPU: AOTAutograd / inductor cache counter mismatch#20Stonepia wants to merge 1 commit into
Stonepia wants to merge 1 commit into
Conversation
[upstream pytorch#3539] [Bug Skip] test_cache_hot_load device=xpu fails on XPU: AOTAutograd / inductor cache counter mismatch
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
[upstream pytorch#3539] [Bug Skip] test_cache_hot_load device=xpu fails on XPU: AOTAutograd / inductor cache counter mismatch
Fixes https://github.com/intel-sandbox/torch-xpu-ops-exp/issues/381
Root Cause: The test exercises the AOTAutograd cache hot-load round trip:
dynamic=True, populate the FX graph cache.torch.compiler.save_cache_artifacts()→ serialise.torch.compiler.load_cache_artifacts()→ deserialise.plus the expected number of
fxgraph_cache_miss/fxgraph_lookup_write_fileevents.On XPU,
autotune_local_cache: Trueproduces extra Triton autotunecache entries (visible as
len(cache_info.autotune_artifacts) == 2where
autotune_expect = 2 if device == GPU_TYPE else 0). Thenon-bundled subclass adds these to the inductor counter
(
Expected 4 but got 6); the bundled subclass appears to bypass theinductor count entirely (
Expected 2 but got 0).This is consistent with the XPU autotune cache pathway not following
the same accounting as the CUDA path, OR the bundled cache
serialisation on XPU not embedding the autotune artifacts into the
bundle the way it does on CUDA.
Failed Tests:
dynamo/test_aot_autograd_cache_xpu.py::AOTAutogradCacheTests::test_cache_hot_load_device_xpu_bfloat16_dynamic_Truedynamo/test_aot_autograd_cache_xpu.py::AOTAutogradCacheBundledTests::test_cache_hot_load_device_xpu_bfloat16_dynamic_Truedynamo/test_aot_autograd_cache_xpu.py::AOTAutogradCacheTests::test_cache_hot_load_device_xpu_float32_dynamic_Truedynamo/test_aot_autograd_cache_xpu.py::AOTAutogradCacheBundledTests::test_cache_hot_load_device_xpu_float32_dynamic_TrueDiff stat: