[Mlir-commits] [flang] [llvm] [mlir] [flang][OpenMP] Enable tiling (PR #143715)

[Jan Leyonberg]

================
@@ -2036,9 +2077,9 @@ static mlir::omp::LoopNestOp genLoopNestOp(
     return llvm::SmallVector<const semantics::Symbol *>(iv);
   };
 
-  auto *nestedEval =
-      getCollapsedLoopEval(eval, getCollapseValue(item->clauses));
-
+  uint64_t nestValue = getCollapseValue(item->clauses);
+  nestValue = nestValue < iv.size() ? iv.size() : nestValue;
----------------
jsjodin wrote:

For tile(2,4) there will be two IVs. The first loop is tiled with size 2, second loop size 4, which will produce 4 loops in the end. If we have collapse(2) and tile(2,3,4) we will have 3 IVs in the omp.loop_nest op.

https://github.com/llvm/llvm-project/pull/143715