[Mlir-commits] [mlir] [mlir][affine][gpu] support unroll dynamic value and apply it to gpu.thread_id op (PR #128113)

[lonely eagle]

================
@@ -884,15 +887,23 @@ void mlir::affine::getTileableBands(
 /// Unrolls this loop completely.
 LogicalResult mlir::affine::loopUnrollFull(AffineForOp forOp) {
   std::optional<uint64_t> mayBeConstantTripCount = getConstantTripCount(forOp);
-  if (mayBeConstantTripCount.has_value()) {
-    uint64_t tripCount = *mayBeConstantTripCount;
-    if (tripCount == 0)
-      return success();
-    if (tripCount == 1)
-      return promoteIfSingleIteration(forOp);
-    return loopUnrollByFactor(forOp, tripCount);
-  }
-  return failure();
+  std::optional<uint64_t> maxMayBeConstantTripCount =
+      getMaxConstantTripCount(forOp);
+
+  if (!mayBeConstantTripCount.has_value() &&
+      !maxMayBeConstantTripCount.has_value())
+    return failure();
+
+  uint64_t tripCount = *mayBeConstantTripCount;
+  uint64_t maxTripCount = *maxMayBeConstantTripCount;
+
+  // Trip equals 0, this loop cannot unroll.
+  if (tripCount <= 0)
+    return success();
+
+  if (tripCount == 1 && maxTripCount == 1)
----------------
linuxlonelyeagle wrote:

If it can run it will definitely be a huge improvement, it's really exciting.

https://github.com/llvm/llvm-project/pull/128113