)]}'
{
  "commit": "f96492221d2bf272053bca6660fc4bdd86592478",
  "tree": "febb2e90b23fec93cb16e323cbaee23f93497d78",
  "parents": [
    "86320e0a8f2d88877775ca685ae3d70ac954faf3"
  ],
  "author": {
    "name": "Kunwar Grover",
    "email": "groverkss@gmail.com",
    "time": "Thu Jul 10 16:11:19 2025 +0100"
  },
  "committer": {
    "name": "GitHub",
    "email": "noreply@github.com",
    "time": "Thu Jul 10 16:11:19 2025 +0100"
  },
  "message": "[mlir][AMDGPU] Add better load/store lowering for full mask (#146748)\n\nThis patch adds a better maskedload/maskedstore lowering on amdgpu\nbackend for loads which are either fully masked or fully unmasked. For\nthese cases, we can either generate a oob buffer load with no if\ncondition, or we can generate a normal load with a if condition (if no\nfat_raw_buffer space).",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "9a368f372c29606bff0e2c84e17407080c316ae9",
      "old_mode": 33188,
      "old_path": "mlir/lib/Dialect/AMDGPU/Transforms/MaskedloadToLoad.cpp",
      "new_id": "60c8660658a953e2a20a6802824bb20a0d7b60ea",
      "new_mode": 33188,
      "new_path": "mlir/lib/Dialect/AMDGPU/Transforms/MaskedloadToLoad.cpp"
    },
    {
      "type": "modify",
      "old_id": "febe46bf7a759a7d47d6036de03a717a7378b75f",
      "old_mode": 33188,
      "old_path": "mlir/test/Dialect/AMDGPU/maskedload-to-load.mlir",
      "new_id": "f1d0ad545539aa6d766f88c0f8ba244ee8388821",
      "new_mode": 33188,
      "new_path": "mlir/test/Dialect/AMDGPU/maskedload-to-load.mlir"
    }
  ]
}