[llvm] [OpenMP] Add pre sm_70 load hack back in (PR #138589)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-offload

Author: Joseph Huber (jhuber6)

<details>
<summary>Changes</summary>

Summary:
Different ordering modes aren't supported for an atomic load, so we just
do an add of zero as the same thing. It's less efficient, but it works.

Fixes https://github.com/llvm/llvm-project/issues/138560


---
Full diff: https://github.com/llvm/llvm-project/pull/138589.diff


1 Files Affected:

- (modified) offload/DeviceRTL/include/Synchronization.h (+4) 


``````````diff
diff --git a/offload/DeviceRTL/include/Synchronization.h b/offload/DeviceRTL/include/Synchronization.h
index f9eb8d0d23198..7e7c8eacb9173 100644
--- a/offload/DeviceRTL/include/Synchronization.h
+++ b/offload/DeviceRTL/include/Synchronization.h
@@ -59,7 +59,11 @@ V add(Ty *Address, V Val, atomic::OrderingTy Ordering,
 template <typename Ty, typename V = utils::remove_addrspace_t<Ty>>
 V load(Ty *Address, atomic::OrderingTy Ordering,
        MemScopeTy MemScope = MemScopeTy::device) {
+#ifdef __NVPTX__
+  return __scoped_atomic_fetch_add(Address, V(0), Ordering, MemScope);
+#else
   return __scoped_atomic_load_n(Address, Ordering, MemScope);
+#endif
 }
 
 template <typename Ty, typename V = utils::remove_addrspace_t<Ty>>

``````````

</details>


https://github.com/llvm/llvm-project/pull/138589