[llvm-branch-commits] [flang] [flang][OpenMP] Map simple `do concurrent` loops to OpenMP host constructs (PR #127633)

[Kareem Ergawy via llvm-branch-commits]

================
@@ -0,0 +1,45 @@
+! Tests that if `do concurrent` is not perfectly nested in its parent loop, that
+! we skip converting the not-perfectly nested `do concurrent` loop.
+
+! RUN: %flang_fc1 -emit-hlfir -fopenmp -fdo-concurrent-to-openmp=host %s -o - \
+! RUN:   | FileCheck %s
+
+program main
+   integer, parameter :: n = 10
+   integer, parameter :: m = 20
+   integer, parameter :: l = 30
+   integer x;
+   integer :: a(n, m, l)
+
+   do concurrent(i=1:n)
+     x = 10
+     do concurrent(j=1:m, k=1:l)
+       a(i,j,k) = i * j + k
+     end do
+   end do
+end
+
+! CHECK: %[[ORIG_K_ALLOC:.*]] = fir.alloca i32 {bindc_name = "k"}
+! CHECK: %[[ORIG_K_DECL:.*]]:2 = hlfir.declare %[[ORIG_K_ALLOC]]
+
+! CHECK: %[[ORIG_J_ALLOC:.*]] = fir.alloca i32 {bindc_name = "j"}
+! CHECK: %[[ORIG_J_DECL:.*]]:2 = hlfir.declare %[[ORIG_J_ALLOC]]
+
+! CHECK: omp.parallel {
+
+! CHECK: omp.wsloop {
+! CHECK: omp.loop_nest ({{[^[:space:]]+}}) {{.*}} {
+! CHECK:   fir.do_loop %[[J_IV:.*]] = {{.*}} {
----------------
ergawy wrote:

I would like to leave it this way to give a better idea of how the whole nested construct was translated.

https://github.com/llvm/llvm-project/pull/127633