[Mlir-commits] [mlir] [MLIR][OpenMP] NFC: Update parallel workshare loop reduction tests (PR #105835)

[Sergio Afonso]

https://github.com/skatrak created https://github.com/llvm/llvm-project/pull/105835

This patch updates MLIR tests for `omp.parallel` + `omp.wsloop` reductions to move the reduction clause into `omp.wsloop` rather than the parent `omp.parallel`, as mandated by the spec for these cases and also to match what Flang is already producing for `parallel do reduction(...)` combined constructs.

>From the OpenMP Spec version 5.2, section 17.2:

> The effect of the reduction clause is as if it is applied to all leaf constructs that permit the clause, except for the following constructs:
>   - The `parallel` construct, when combined with the `sections`, worksharing-loop, `loop`, or `taskloop` construct; [...]

>From 3e443347f2bed5cf4035446bf6e7427360c4a3c3 Mon Sep 17 00:00:00 2001
From: Sergio Afonso <safonsof at amd.com>
Date: Fri, 23 Aug 2024 14:58:21 +0100
Subject: [PATCH] [MLIR][OpenMP] NFC: Update parallel workshare loop reduction
 tests

This patch updates MLIR tests for `omp.parallel` + `omp.wsloop` reductions to
move the reduction clause into `omp.wsloop` rather than the parent
`omp.parallel`, as mandated by the spec for these cases and also to match what
Flang is already producing for `parallel do reduction(...)` combined
constructs.

>From the OpenMP Spec version 5.2, section 17.2:

> The effect of the reduction clause is as if it is applied to all leaf constructs that permit the clause, except for the following constructs:
>   - The `parallel` construct, when combined with the `sections`, worksharing-loop, `loop`, or `taskloop` construct; [...]
---
 mlir/test/Dialect/OpenMP/ops.mlir             | 16 ++++++++--------
 mlir/test/Target/LLVMIR/openmp-reduction.mlir |  4 ++--
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/mlir/test/Dialect/OpenMP/ops.mlir b/mlir/test/Dialect/OpenMP/ops.mlir
index 9c308cc0108493..7ba55fc957a473 100644
--- a/mlir/test/Dialect/OpenMP/ops.mlir
+++ b/mlir/test/Dialect/OpenMP/ops.mlir
@@ -1019,10 +1019,10 @@ func.func @parallel_reduction_byref() {
 func.func @parallel_wsloop_reduction(%lb : index, %ub : index, %step : index) {
   %c1 = arith.constant 1 : i32
   %0 = llvm.alloca %c1 x i32 : (i32) -> !llvm.ptr
-  // CHECK: omp.parallel reduction(@add_f32 %{{.*}} -> %{{.+}} : !llvm.ptr) {
-  omp.parallel reduction(@add_f32 %0 -> %prv : !llvm.ptr) {
-    // CHECK: omp.wsloop {
-    omp.wsloop {
+  // CHECK: omp.parallel {
+  omp.parallel {
+    // CHECK: omp.wsloop reduction(@add_f32 %{{.*}} -> %{{.+}} : !llvm.ptr) {
+    omp.wsloop reduction(@add_f32 %0 -> %prv : !llvm.ptr) {
       // CHECK: omp.loop_nest (%{{.+}}) : index = (%{{.+}}) to (%{{.+}}) step (%{{.+}}) {
       omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
         %1 = arith.constant 2.0 : f32
@@ -1216,10 +1216,10 @@ func.func @parallel_reduction2() {
 func.func @parallel_wsloop_reduction2(%lb : index, %ub : index, %step : index) {
   %c1 = arith.constant 1 : i32
   %0 = llvm.alloca %c1 x i32 : (i32) -> !llvm.ptr
-  // CHECK: omp.parallel reduction(@add2_f32 %{{.*}} -> %{{.+}} : !llvm.ptr) {
-  omp.parallel reduction(@add2_f32 %0 -> %prv : !llvm.ptr) {
-    // CHECK: omp.wsloop {
-    omp.wsloop {
+  // CHECK: omp.parallel {
+  omp.parallel {
+    // CHECK: omp.wsloop reduction(@add2_f32 %{{.*}} -> %{{.+}} : !llvm.ptr) {
+    omp.wsloop reduction(@add2_f32 %0 -> %prv : !llvm.ptr) {
       // CHECK: omp.loop_nest (%{{.+}}) : index = (%{{.+}}) to (%{{.+}}) step (%{{.+}}) {
       omp.loop_nest (%iv) : index = (%lb) to (%ub) step (%step) {
         %1 = arith.constant 2.0 : f32
diff --git a/mlir/test/Target/LLVMIR/openmp-reduction.mlir b/mlir/test/Target/LLVMIR/openmp-reduction.mlir
index bfdad8c19335e1..1d4b4915bcc399 100644
--- a/mlir/test/Target/LLVMIR/openmp-reduction.mlir
+++ b/mlir/test/Target/LLVMIR/openmp-reduction.mlir
@@ -540,8 +540,8 @@ llvm.func @parallel_nested_workshare_reduction(%ub : i64) {
   %lb = llvm.mlir.constant(1 : i64) : i64
   %step = llvm.mlir.constant(1 : i64) : i64
   
-  omp.parallel reduction(@add_i32 %0 -> %prv : !llvm.ptr) {
-    omp.wsloop {
+  omp.parallel {
+    omp.wsloop reduction(@add_i32 %0 -> %prv : !llvm.ptr) {
       omp.loop_nest (%iv) : i64 = (%lb) to (%ub) step (%step) {
         %ival = llvm.trunc %iv : i64 to i32
         %lprv = llvm.load %prv : !llvm.ptr -> i32