From c6e0f416e7aafeef473839e4a24517fe0e5909dc Mon Sep 17 00:00:00 2001
From: Ahmed <36049290+AhmedAmraniAkdi@users.noreply.github.com>
Date: Tue, 28 Oct 2025 16:04:51 +0000
Subject: [PATCH] Specify memory scope and memory order for the atomic
 operations in generic_address_space generic_atomics_variant
 generic_atomics_invariant (#2550)

Use the explicit version of the atomic_load/store and atomic_fetch_add
with memory order relaxed and memory scope workgroup to allow devices
that only support the minimum CL_DEVICE_ATOMIC_MEMORY_CAPABILITIES which
are (CL_DEVICE_ATOMIC_ORDER_RELAXED | CL_DEVICE_ATOMIC_SCOPE_WORK_GROUP)
to run the tests.

The test should only require the relaxed ordering and memory scope
workgroup anyway.

From the specificiation:

"The non-explicit atomic_store function requires support for OpenCL C
2.0, or OpenCL C 3.0 or newer and both the
__opencl_c_atomic_order_seq_cst and __opencl_c_atomic_scope_device
features"

"The non-explicit atomic_load function requires support for OpenCL C 2.0
or OpenCL C 3.0 or newer and both the __opencl_c_atomic_order_seq_cst
and __opencl_c_atomic_scope_device features."

"The non-explicit atomic_fetch_key functions require support for OpenCL
C 2.0, or OpenCL C 3.0 or newer and both the
__opencl_c_atomic_order_seq_cst and __opencl_c_atomic_scope_device
features."
---
 .../generic_address_space/atomic_tests.cpp        | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/test_conformance/generic_address_space/atomic_tests.cpp b/test_conformance/generic_address_space/atomic_tests.cpp
index 8a568f0a..7ab65d36 100644
--- a/test_conformance/generic_address_space/atomic_tests.cpp
+++ b/test_conformance/generic_address_space/atomic_tests.cpp
@@ -35,7 +35,7 @@ kernel void testKernel(global atomic_int* globalPtr, local atomic_int* localPtr)
     int wgid = get_group_id(0);
     int wgsize = get_local_size(0);
 
-    if (tid == 0) atomic_store(localPtr, 0);
+    if (tid == 0) atomic_store_explicit(localPtr, 0, memory_order_relaxed, memory_scope_work_group);
 
     barrier(CLK_LOCAL_MEM_FENCE);
 
@@ -47,12 +47,12 @@ kernel void testKernel(global atomic_int* globalPtr, local atomic_int* localPtr)
     if ((wgid % 2) == 0)
         ptr = localPtr;
 
-    int inc = atomic_fetch_add(ptr, 1);
+    int inc = atomic_fetch_add_explicit(ptr, 1, memory_order_relaxed, memory_scope_work_group);
 
     // In the cases where the local memory ptr was used,
     // save off the final value.
     if ((wgid % 2) == 0 && inc == (wgsize-1))
-        atomic_store(&globalPtr[wgid], inc);
+        atomic_store_explicit(&globalPtr[wgid], inc, memory_order_relaxed, memory_scope_work_group);
 }
 )OpenCLC";
 
@@ -67,7 +67,7 @@ kernel void testKernel(global atomic_int* globalPtr, local atomic_int* localPtr)
     int wgid = get_group_id(0);
     int wgsize = get_local_size(0);
 
-    if (tid == 0) atomic_store(localPtr, 0);
+    if (tid == 0) atomic_store_explicit(localPtr, 0, memory_order_relaxed, memory_scope_work_group);
 
     barrier(CLK_LOCAL_MEM_FENCE);
 
@@ -79,14 +79,17 @@ kernel void testKernel(global atomic_int* globalPtr, local atomic_int* localPtr)
     if ((tid % 2) == 0)
         ptr = localPtr;
 
-    atomic_fetch_add(ptr, 1);
+    atomic_fetch_add_explicit(ptr, 1, memory_order_relaxed, memory_scope_work_group);
 
     barrier(CLK_LOCAL_MEM_FENCE);
 
     // In the cases where the local memory ptr was used,
     // save off the final value.
     if (tid == 0)
-        atomic_store(&globalPtr[(wgid * 2) + 1], atomic_load(localPtr));
+        atomic_store_explicit(&globalPtr[(wgid * 2) + 1],
+                                atomic_load_explicit(localPtr, memory_order_relaxed, memory_scope_work_group),
+                                memory_order_relaxed,
+                                memory_scope_work_group);
 }
 )OpenCLC";
 }