[user-streams] Test inductor streams infra by mlazos · Pull Request #174224 · pytorch/pytorch

mlazos · 2026-02-03T18:43:19Z

Stack from ghstack (oldest at bottom):

Add comprehensive tests for user stream support in inductor:

Stream utility tests (pool, context manager, naming)
Event factory tests (creation, ordering, hashing)
Wrapper codegen tests (stream context enter/exit)
Compile tests for stream semantics preservation

The compile tests verify that torch.compile() correctly handles:

Stream context managers
Event record/wait operations
Multi-stream synchronization patterns
Fusion behavior within and across streams

Test assertions check for generated code patterns that may appear as
either custom ops (record_event/wait_event) or method calls.

cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben @jataylo

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. [ghstack-poisoned]

pytorch-bot · 2026-02-03T18:43:27Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/174224

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (3 Unrelated Failures)

As of commit c4f3ef3 with merge base d2d12ef ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

inductor / unit-test / inductor-test / test (inductor, 1, 2, linux.g5.4xlarge.nvidia.gpu) (gh) (similar failure)
test/inductor/test_torchinductor_opinfo.py::TestInductorOpInfoCUDA::test_comprehensive_pca_lowrank_cuda_float32

BROKEN TRUNK - The following job failed but was present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

trunk / macos-py3-arm64 / test (default, 3, 3, macos-m1-stable) (gh) (trunk failure)
test/test_sparse.py::TestSparseMPS::test_is_nonzero_mps

UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:

inductor / inductor-cpu-test / test (cpu_inductor_torchbench, 1, 2, linux.2xlarge.amx, unstable) (gh) (#174929)
detectron2_maskrcnn_r_50_fpn

This comment was automatically generated by Dr. CI and updates every 15 minutes.

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. ghstack-source-id: c4b1c07 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. ghstack-source-id: 05f6722 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. ghstack-source-id: 05f6722 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. ghstack-source-id: 3349dba Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. ghstack-source-id: 8083c86 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. ghstack-source-id: 012339c Pull Request resolved: #174224

ghstack-source-id: e3bfed4c3b2c8f3b9a0e7d5c1a4f2b6d8c0a9e7f Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ghstack-source-id: 947320e Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ghstack-source-id: 9594123 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ghstack-source-id: 9594123 Pull Request resolved: #174224

mlazos · 2026-03-11T18:07:12Z

@claude review this pr

claude · 2026-03-11T18:07:44Z

Claude finished @mlazos's task in 3m 16s —— View job

Review of PR #174224 — `[user-streams] Test inductor streams infra`

Read the test file and gather context
Review code quality and correctness
Check adherence to PyTorch testing conventions
Post review feedback

Overall this is a solid and comprehensive test suite for the user stream infrastructure. The test coverage spans unit tests for constants/utilities, codegen verification, and end-to-end compilation tests with increasingly complex stream topologies. A few items to address:

Style: Remove docstrings from test methods

The existing inductor test files (test_cuda_repro.py, test_triton_kernels.py, test_torchinductor.py) do not use docstrings on test methods — they rely on descriptive method names instead. Every test method in this file has a docstring, which is inconsistent with the codebase convention. The module-level docstring (lines 2-6) is fine, but the per-method ones should be removed.

Fix this →

Style: Move `run_and_get_code` import to module level

from torch._inductor.utils import run_and_get_code is imported inside every test method in TestUserStreamCompile and TestGenericStreamCompile. The convention in the inductor test suite is to import this at the module level (see test_torchinductor.py:57, test_cuda_repro.py:24).

Fix this →

Unnecessary `instantiate_parametrized_tests` calls

instantiate_parametrized_tests is called on all five test classes (lines 1061-1065), but none of them use @parametrize on any method. While this is technically harmless, it adds noise. If you plan to add parametrized tests later, this is fine to leave — but if not, it would be cleaner to remove these calls and the corresponding import.

Weak assertion: `test_event_wait_without_record` (line 401)

The test name says "Test that waiting on unrecorded event works (no-op)" but the test body actually does record the event before waiting (line 410-411: event.record()). The test name and docstring are misleading — consider renaming to something like test_event_record_then_wait or fixing the test body to match what it claims to test.

Weak code assertions in `test_multiple_stream_contexts` and similar

Several tests use a fallback pattern like:

self.assertTrue(
    code.count("torch.cuda.stream") >= 1 or "stream" in code.lower(),
    "Expected stream context in generated code",
)

The second condition ("stream" in code.lower()) is extremely permissive — virtually any generated code referencing variables like default_stream would pass this. This makes these assertions nearly vacuous. If the intent is to allow for scheduler optimizations that might eliminate explicit stream contexts, consider either:

Asserting something more specific about the generated code structure, or
Documenting explicitly what optimizations could cause torch.cuda.stream to be absent

This pattern appears at lines: 208-211, 241-244, 559-562, 612-615, 718-721, 804-807.

`test_no_fusion_across_streams`: kernel count assertion may be fragile

triton_kernel_count = code.count("triton_") + code.count(".run(")
self.assertGreaterEqual(triton_kernel_count, 2, ...)

Counting "triton_" and ".run(" substrings is fragile — "triton_" could appear in comments, import statements, or variable names unrelated to kernel calls; ".run(" could match non-kernel method calls. Consider using torch._inductor.metrics or a more structured approach (e.g., torch._inductor.utils.run_and_get_code already structures the output). This applies to lines 811 and 861.

Inconsistent CUDA skip decorators

TestUserStreamCompile uses @unittest.skipIf(not TEST_CUDA, "requires CUDA") (line 116), while TestGenericStreamCompile uses @unittest.skipUnless(TEST_CUDA, "requires CUDA") (line 900). Both are functionally equivalent, but using different forms is inconsistent. The inductor test suite commonly uses @requires_gpu or @requires_cuda from torch.testing._internal.inductor_utils — consider aligning with those.

Missing `test_fusion_within_same_stream` assertion

In test_fusion_within_same_stream (line 868), there's a comment at line 896-898 saying these ops "should be fused into a single kernel" but there's no assertion verifying this. The test only checks correctness and that torch.cuda.stream is in the code. Consider adding an assertion on kernel count (e.g., assertEqual(code.count("triton_poi"), 1) or similar) to actually verify the fusion claim.

Minor: `test_exit_cuda_stream_context` test is nearly a no-op

def test_exit_cuda_stream_context(self):
    line = ExitCudaStreamContextLine()
    self.assertIsNotNone(line)

This only tests that the constructor doesn't crash. Since ExitCudaStreamContextLine is a @dataclass with no fields, there's not much to test here. Consider either removing this test or exercising the codegen() method (as done in TestStreamCodegen.test_exit_cuda_stream_context_codegen).

Summary

The test file provides good coverage of the stream infrastructure at multiple levels (unit, codegen, integration). The main issues are:

Style nits — docstrings on every method, local imports, inconsistent skip decorators
Weak assertions — several codegen assertions are too permissive to catch regressions
One misleading test name (test_event_wait_without_record)
Missing fusion assertion in test_fusion_within_same_stream

None of these are blocking, but addressing them would strengthen the test suite and align with the codebase conventions.

test/inductor/test_user_streams.py

ghstack-source-id: 9594123 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ghstack-source-id: 9bbdf58 Pull Request resolved: #174224

ghstack-source-id: 9594123 Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

ghstack-source-id: 8d33e1a Pull Request resolved: #174224

Add comprehensive tests for user stream support in inductor: - Stream utility tests (pool, context manager, naming) - Event factory tests (creation, ordering, hashing) - Wrapper codegen tests (stream context enter/exit) - Compile tests for stream semantics preservation The compile tests verify that torch.compile() correctly handles: - Stream context managers - Event record/wait operations - Multi-stream synchronization patterns - Fusion behavior within and across streams Test assertions check for generated code patterns that may appear as either custom ops (record_event/wait_event) or method calls. cc voznesenskym penguinwu EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng wenzhe-nrv jiayisunx ipiszy kadeng muchulee8 amjames chauhang aakhundov coconutruben jataylo [ghstack-poisoned]

Add kernel count verification using generated_kernel_count and FX graph snapshots via CompileCounterWithBackend/normalize_gm to the stream fusion and codegen structure tests. Authored with Claude. ghstack-source-id: 299aec2 Pull Request resolved: #174224

shunting314 · 2026-03-13T17:54:54Z

test/inductor/test_user_streams.py

+        self.assertExpectedInline(
+            normalize_gm(counter.graphs[0].print_readable(print_output=False)),
+            """\
+class GraphModule(torch.nn.Module):


Would these string matching being fragile and may get broken often by future unrelated change?

mlazos · 2026-03-13T18:28:41Z

@pytorchbot merge

pytorchmergebot · 2026-03-13T18:30:58Z

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

pytorch-bot bot added ciflow/inductor module: inductor topic: not user facing topic category labels Feb 3, 2026

mlazos added a commit that referenced this pull request Mar 3, 2026

[user-streams] Test inductor streams infra

06687d7

ghstack-source-id: e3bfed4c3b2c8f3b9a0e7d5c1a4f2b6d8c0a9e7f Pull Request resolved: #174224

mlazos added a commit that referenced this pull request Mar 3, 2026

[user-streams] Test inductor streams infra

435e6b5

ghstack-source-id: 947320e Pull Request resolved: #174224

mlazos added 2 commits March 3, 2026 14:03

This was referenced Mar 6, 2026

[user-streams] Remove cross-stream dependency codegen #176695

Closed

[user-streams] Add stream support to scheduler #176696

Closed

mlazos added a commit that referenced this pull request Mar 6, 2026

[user-streams] Test inductor streams infra

eeef6ad

ghstack-source-id: 9594123 Pull Request resolved: #174224

mlazos mentioned this pull request Mar 6, 2026

[user-streams] Remove cross-stream dependency codegen #176700

Closed

mlazos added a commit that referenced this pull request Mar 6, 2026

[user-streams] Test inductor streams infra

79dade6

ghstack-source-id: 9594123 Pull Request resolved: #174224

mlazos requested a review from eellison March 6, 2026 18:07

mlazos added a commit that referenced this pull request Mar 6, 2026

[user-streams] Test inductor streams infra

e169e59

ghstack-source-id: 9594123 Pull Request resolved: #174224

mlazos added a commit that referenced this pull request Mar 10, 2026

[user-streams] Test inductor streams infra

1fcd23e

ghstack-source-id: 9594123 Pull Request resolved: #174224

shunting314 reviewed Mar 11, 2026

View reviewed changes

test/inductor/test_user_streams.py Outdated Show resolved Hide resolved

test/inductor/test_user_streams.py Show resolved Hide resolved

mlazos added a commit that referenced this pull request Mar 11, 2026

[user-streams] Test inductor streams infra

523b557

ghstack-source-id: 9594123 Pull Request resolved: #174224

mlazos added a commit that referenced this pull request Mar 11, 2026

[user-streams] Test inductor streams infra

ceb5651

ghstack-source-id: 9bbdf58 Pull Request resolved: #174224

mlazos added a commit that referenced this pull request Mar 11, 2026

[user-streams] Test inductor streams infra

b57e3fb

ghstack-source-id: 9594123 Pull Request resolved: #174224

mlazos added a commit that referenced this pull request Mar 11, 2026

[user-streams] Test inductor streams infra

5d24675

ghstack-source-id: 8d33e1a Pull Request resolved: #174224

mlazos requested a review from shunting314 March 12, 2026 22:58

mlazos added the ciflow/trunk Trigger trunk jobs on your pull request label Mar 13, 2026

shunting314 approved these changes Mar 13, 2026

View reviewed changes

pytorchmergebot added the merging label Mar 13, 2026

pytorchmergebot added the Merged label Mar 13, 2026

pytorchmergebot closed this in 886deed Mar 13, 2026

pytorchmergebot removed the merging label Mar 13, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[user-streams] Test inductor streams infra#174224

[user-streams] Test inductor streams infra#174224
mlazos wants to merge 21 commits intogh/mlazos/94/basefrom
gh/mlazos/94/head

mlazos commented Feb 3, 2026 •

edited

Loading

Uh oh!

pytorch-bot bot commented Feb 3, 2026 •

edited

Loading

Uh oh!

mlazos commented Mar 11, 2026

Uh oh!

claude bot commented Mar 11, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

shunting314 Mar 13, 2026

Uh oh!

mlazos commented Mar 13, 2026

Uh oh!

pytorchmergebot commented Mar 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

mlazos commented Feb 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Feb 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/174224

✅ You can merge normally! (3 Unrelated Failures)

Uh oh!

mlazos commented Mar 11, 2026

Uh oh!

claude bot commented Mar 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review of PR #174224 — [user-streams] Test inductor streams infra

Style: Remove docstrings from test methods

Style: Move run_and_get_code import to module level

Unnecessary instantiate_parametrized_tests calls

Weak assertion: test_event_wait_without_record (line 401)

Weak code assertions in test_multiple_stream_contexts and similar

test_no_fusion_across_streams: kernel count assertion may be fragile

Inconsistent CUDA skip decorators

Missing test_fusion_within_same_stream assertion

Minor: test_exit_cuda_stream_context test is nearly a no-op

Summary

Uh oh!

Uh oh!

Uh oh!

shunting314 Mar 13, 2026

Choose a reason for hiding this comment

Uh oh!

mlazos commented Mar 13, 2026

Uh oh!

pytorchmergebot commented Mar 13, 2026

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

mlazos commented Feb 3, 2026 •

edited

Loading

pytorch-bot bot commented Feb 3, 2026 •

edited

Loading

claude bot commented Mar 11, 2026 •

edited

Loading

Review of PR #174224 — `[user-streams] Test inductor streams infra`

Style: Move `run_and_get_code` import to module level

Unnecessary `instantiate_parametrized_tests` calls

Weak assertion: `test_event_wait_without_record` (line 401)

Weak code assertions in `test_multiple_stream_contexts` and similar

`test_no_fusion_across_streams`: kernel count assertion may be fragile

Missing `test_fusion_within_same_stream` assertion

Minor: `test_exit_cuda_stream_context` test is nearly a no-op