[Mlir-commits] [mlir] [mlir][SCF] Allow using a custom operation to generate loops with `mlir::tileUsingSCF`. (PR #159660)

[Quinn Dawkins]

================
@@ -117,6 +121,98 @@ struct SCFTilingOptions {
     reductionDims.insert(dims.begin(), dims.end());
     return *this;
   }
+
+  //-------------------------------------------------------------------------//
+  // Options related to tiling using custom loop.
+  //-------------------------------------------------------------------------//
+
+  // For generating the inter-tile loops using a custom loop, two callback
+  // functions are needed
+  // 1. That generates the "loop header", i.e. the loop that iterates over the
+  //    different tiles.
+  // 2. That generates the loop terminator
+  //
+  // For `scf.forall` case the call back to generate loop header would generate
+  //
+  // ```mlir
+  // scf.forall (...) = ... {
+  //   ..
+  // }
+  // ```
+  //
+  // and the call back to generate the loop terminator would generate the
+  // `scf.in_parallel` region
+  //
+  // ```mlir
+  // scf.forall (...) = ... {
+  //   scf.in_parallel {
+  //      tensor.parallel_insert_slice ...
+  //   }
+  // }
+  // ```
+  //
+
+  // Information that is to be returned by the callback to generate the loop
+  // header needed for the rest of the tiled codegeneration.
----------------
qedawkins wrote:

Had to read this sentence a few times to parse it. Maybe this
```suggestion
  // Information returned by the loop header callback
  // needed for the rest of the tiled codegeneration.
```

https://github.com/llvm/llvm-project/pull/159660