Enable TensorIndexer with the transpose tests

[naoyam]

Enabling TensorIndexer for the transpose tests. Diff check revealed some mismatches, but I believe they are benign. As far as I can see, all of them are caused by some slight difference for issue #681. TensorIndexer's workaround is simpler but not as precise as the legacy indexer, as indicated in #2689.

For example, there is an index for a predicate of TransposeTest.FusionScheduleTransposeSimple:

 ( ( ( blockIdx.x % ( ( ceilDiv(( (( (( getMetaData(T0) )).logical_size ))[2] ), 32) ) * ( ceilDiv(( (( (( getMetaData(T0) )).logical_size ))[1] ), 32) ) ) ) / ( ceilDiv(( (( (( getMetaData(T0) )).logical_size ))[1] ), 32) ) ) * 32 ) + (threadIdx.x * 4) % 32)

The above is generated by the legacy indexer. With TensorIndexer, it's generated as:

( ( ( ( blockIdx.x % ( ( ceilDiv(( (( (( getMetaData(T0) )).logical_size ))[2] ), 32) ) * ( ceilDiv(( (( (( getMetaData(T0) )).logical_size ))[1] ), 32) ) ) ) / ( ceilDiv(( (( (( getMetaData(T0) )).logical_size ))[1] ), 32) ) ) * 32 ) + 0 )

The difference is threadIdx.x 4 % 32 vs. 0. Note that this is a predicate for the condition at the start position, i.e., they are predicated with >= 0, and as such the index of TensorIndexer is more conservative, so the legacy indexer may have a higher chance of using the fast path created by the unswitch predicate. However, in practice, both predicates should be always true, so it shouldn't matter. I don't see any actual difference with stop predicates.

[naoyam]

!test --diff

[github-actions]

Review updated until commit b28d962

Description

• Enable TensorIndexer with transpose tests

• Add SetUp method to configure IdModel options

---

Changes walkthrough 📝

	Relevant files
Enhancement	test_gpu_transpose.cpp Add SetUp for TensorIndexer configuration                                tests/cpp/test_gpu_transpose.cpp Added SetUp method to override NVFuserTest::SetUp Configured IdModel options in SetUp method +5/-0

---

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

🧪 PR contains tests

⚡ Recommended focus areas for review Option Setting The EnableOption::IdModel is set to "all" in the SetUp method. This might have unintended side effects or could be too broad. It would be good to verify if this is necessary and if there are any specific options that should be enabled instead. NVFuserTest::SetUp(); EnableOptionsGuard::getCurOptions().set(EnableOption::IdModel, {"all"}); }

[naoyam]

!test --diff

[naoyam]

!test --diff

[naoyam]

!test --diff

[naoyam]

!test --diff

[greptile-apps]

Greptile Summary

• Enables TensorIndexer for all transpose tests by setting IdModel option to "all" in SetUp() method
• Tests previously used legacy indexer which generates more precise but complex predicates for edge cases (issue Wrong unswitch predicate  #681)

Confidence Score: 5/5

• Safe to merge - enables TensorIndexer for existing test suite with benign differences explained
• This PR only modifies test configuration to enable TensorIndexer. The author has verified all tests pass and explained that minor predicate differences are benign and more conservative. No production code changes.
• No files require special attention

Important Files Changed

Filename	Overview
tests/cpp/test_transpose.cpp	Added SetUp() override to enable TensorIndexer (IdModel with "all" option) for all transpose tests

Sequence Diagram

sequenceDiagram
    participant Test as TransposeTest
    participant Setup as SetUp()
    participant Options as EnableOptionsGuard
    participant IdModel as TensorIndexer
    participant Tests as Test Cases
    
    Test->>Setup: "Initialize test fixture"
    Setup->>Options: "set(EnableOption::IdModel, {\"all\"})"
    Options->>IdModel: "Enable TensorIndexer for all operations"
    Setup->>Tests: "Run transpose tests"
    Tests->>IdModel: "Generate indices using TensorIndexer"
    IdModel->>Tests: "Return conservative predicates"

Loading

[greptile-apps]

_{1 file reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile
React with 👍 or 👎 to share your feedback on this new summary format}

[jjsjann123]

LGTM