platform: Clarify behavior for ATOMIC_SCOPE_ALL_DEVICES #1129

Clarify behavior of above capability value on devices that don't support fine grained SVM
KhronosGroup · May 21, 2024 · ea27dd6 · ea27dd6
1 parent 0483627
commit ea27dd6
Showing 1 changed file with 2 additions and 0 deletions.
diff --git a/api/opencl_platform_layer.asciidoc b/api/opencl_platform_layer.asciidoc
@@ -1606,6 +1606,8 @@ include::{generated}/api/version-notes/CL_DEVICE_ATOMIC_MEMORY_CAPABILITIES.asci
  {CL_DEVICE_ATOMIC_ORDER_RELAXED} \| +
  {CL_DEVICE_ATOMIC_SCOPE_WORK_GROUP}
 
+ A device that only supports {CL_DEVICE_SVM_COARSE_GRAIN_BUFFER_anchor} as per {CL_DEVICE_SVM_CAPABILITIES_anchor} query can still report support for {CL_DEVICE_ATOMIC_SCOPE_ALL_DEVICES_anchor} in queries for {CL_DEVICE_ATOMIC_MEMORY_CAPABILITIES_anchor} and {CL_DEVICE_ATOMIC_FENCE_CAPABILITIES_anchor}. Operations parameterized with memory_scope_all_svm_devices on such a device will behave as if they were parameterized with memory_scope_device.
+
 | {CL_DEVICE_ATOMIC_FENCE_CAPABILITIES_anchor}
 
 include::{generated}/api/version-notes/CL_DEVICE_ATOMIC_FENCE_CAPABILITIES.asciidoc[]