Use a much simpler guard for version 1.6, which allows for direct CUDA indexing to be used.

[ROCm/hip commit: 7acb1e6ff6]

projects/hip/include/hip/hcc_detail/host_defines.h

@@ -44,8 +44,12 @@ THE SOFTWARE.
 #if GENERIC_GRID_LAUNCH == 0
 #define __global__  __attribute__((hc_grid_launch)) __attribute__((used))
 #else
-#define __global__ \
-    __attribute__((annotate("__HIP_global_function__"), cpu, hc, used))
+    #if __hcc_workweek__ >=17481
+        #define __global__ \
+            __attribute__((annotate("__HIP_global_function__"), cpu, hc, used))
+    #else
+        #define __global__ __attribute__((hc, used))
+    #endif
 #endif //GENERIC_GRID_LAUNCH
 
 #define __noinline__      __attribute__((noinline))