[Mlir-commits] [mlir] [MLIR] [SparseTensor] Implement multiple loop ordering heuristics for sparse tensor dialect (PR #151885)

[Aart Bik]

================
@@ -55,6 +55,33 @@ enum class SparseEmitStrategy {
   kDebugInterface, // generate only place-holder for sparse iteration
 };
 
+namespace sparse_tensor {
+/// Select between different loop ordering strategies.
+/// Loop ordering strategies for sparse tensor compilation.
+/// These strategies control how loops are ordered during sparsification,
+/// providing 3-71% performance improvements across diverse workloads.
+enum class LoopOrderingStrategy : unsigned {
+  kDefault,        ///< Default: Prefer parallel loops to reduction loops.
----------------
aartbik wrote:

What is this ///< style

(I am not familiar with that, apologies if this is standard)

https://github.com/llvm/llvm-project/pull/151885