[clang] [CIR][OpenMP] Enable lowering of the OpenMP dialect to LLVM IR (PR #178515)

[Jan Leyonberg via cfe-commits]

================
@@ -23,6 +23,8 @@ void emit_simple_parallel() {
     during(i);
   }
   // CHECK-NEXT: omp.parallel {
+  // CHECK-NEXT: {{.*}} = cir.load align(4) %{{.*}} : !cir.ptr<!s32i>, !s32i
----------------
jsjodin wrote:

When you say LLVM, do you mean a more extensive test that just checking for the OMP runtime calls? I can do that for both this and OGCG if you like. I looked at the differences and the code comes out pretty different, but looks equivalent.
I was able to run another simple test initializing an array with omp_get_thread_num() and printing it out and it was giving different values in the array so runtime wise it seems to work. Do we also add runtime tests somewhere?

https://github.com/llvm/llvm-project/pull/178515