)]}'
{
  "commit": "8823fe65dcfa6f95985b14be5db40b1d86edbe10",
  "tree": "2f8da8037704372e65162c7545087e0f61e28c89",
  "parents": [
    "887e0b6c143e58f95e5b748e4d2919c26bbb5ca1"
  ],
  "author": {
    "name": "ziereis",
    "email": "44057120+ziereis@users.noreply.github.com",
    "time": "Tue Oct 06 13:07:07 2026 +0200"
  },
  "committer": {
    "name": "GitHub",
    "email": "noreply@github.com",
    "time": "Tue Oct 06 13:07:07 2026 +0200"
  },
  "message": "[Integrate] Update torch-mlir to 528d3dcb (#24908)\n\nUpdates torch-mlir from its July 7 upstream baseline to\n[`llvm/torch-mlir@528d3dcb`](https://github.com/llvm/torch-mlir/commit/528d3dcb9a2288743b62df9f4937b1689787e38a),\nusing\n[`iree-org/torch-mlir@137b8fb8`](https://github.com/iree-org/torch-mlir/commit/137b8fb8596f4c8091f69535e5153f35b7ece311)\n(`sm-iree-integrates/torch-20261002`) adapted to IREE\u0027s current LLVM\nrevision.\n\nOn top of canonical torch-mlir, the integration branch carries the\nadaptation patches already on IREE\u0027s current pin that upstream does not\nhave yet: the LLVM submodule bump, `SymbolOpInterface` for\n`ClassTypeOp`/`GlobalSlotOp`, and the TMTensor strict-properties\nassembly fix. The `DenseMapInfo` and `RegionSuccessor` adaptations from\nthe previous pin are now upstream and were dropped.\n\nONNX op test expectations:\n\n- Now passing, removed from `expected_run_failures` in the CI-exercised\nconfigs:\n`averagepool_3d_dilations_large_count_include_pad_is_0_ceil_mode_is_True`,\n`nonmaxsuppression_flipped_coordinates`, `reduce_{max,min}_bool_inputs`.\n\n\n- `test_reversesequence_time` now fails to compile on llvm-cpu and is\nmarked as an expected compile failure. ReverseSequence now lowers to an\n`scf.for` over the time axis, so each dispatch reverses a column picked\nby a dynamic offset. The resulting `vector.gather` reads a strided\ncolumn subview with a dynamic offset, which neither\n`RemoveStrideFromGatherSource` (bails on dynamic offsets) nor\n`Gather1DToConditionalLoads` (rejects non-unit inner stride) handles, so\nthe gather reaches LLVM conversion unlowered. The proper fix is\ndynamic-offset support in the upstream MLIR gather lowering.\n\n🤖 Generated with [Claude Code](https://claude.com/claude-code)\n\nSigned-off-by: Thomas Ziereis \u003cziereis@roofline.ai\u003e",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "b91d549184484e2168c68acf9ae68388264fb55c",
      "old_mode": 33188,
      "old_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O0.json",
      "new_id": "c8d9bf1dcbeacd39a0e080ff9d66c4f0d0d80a0f",
      "new_mode": 33188,
      "new_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O0.json"
    },
    {
      "type": "modify",
      "old_id": "a3055a7a504ed4413410d9a89b8b5279c7a8ccbe",
      "old_mode": 33188,
      "old_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O2.json",
      "new_id": "7b1552719e94a0b477b38cd4122166c39ee82ba6",
      "new_mode": 33188,
      "new_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O2.json"
    },
    {
      "type": "modify",
      "old_id": "e03c849580766688c12516cabe6d9d0dfb47c401",
      "old_mode": 33188,
      "old_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O3.json",
      "new_id": "d5174187c4e4c49fdb53ba7d1c4632eb201b6532",
      "new_mode": 33188,
      "new_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O3.json"
    },
    {
      "type": "modify",
      "old_id": "8744f2aa28ef0aff20bf48c3f0d72785a7067125",
      "old_mode": 33188,
      "old_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna4_O3.json",
      "new_id": "0238939f1a83208bd550bbd37cbe043966f46c27",
      "new_mode": 33188,
      "new_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna4_O3.json"
    },
    {
      "type": "modify",
      "old_id": "5ccfb663144ced86ec97602261cde8803768d445",
      "old_mode": 33188,
      "old_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_vulkan_rdna3_O0.json",
      "new_id": "26a6f9ba7595d9ee1b11e66297da1733268cbf2c",
      "new_mode": 33188,
      "new_path": "tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_vulkan_rdna3_O0.json"
    },
    {
      "type": "modify",
      "old_id": "679f0e3894a61fae69168728c0f08a1b67819ca8",
      "old_mode": 57344,
      "old_path": "third_party/torch-mlir",
      "new_id": "137b8fb8596f4c8091f69535e5153f35b7ece311",
      "new_mode": 57344,
      "new_path": "third_party/torch-mlir"
    }
  ]
}